UniMitra Skills
Industry-grade projects
Portfolio projects on real, licensed data and real systems. Each brief lists its datasets and their licences, the deliverables and how a reviewer would grade it.
- DataData AnalystTurn messy business data into answers people act on, with SQL, a BI tool and clear writing.24 to 40 weeks77 lessons12 projects
- DataData ScientistBuild models and experiments that change decisions, and explain them so people trust them.32 to 52 weeks63 lessons12 projects
- DataData EngineerBuild the pipelines, warehouses and streams that every analyst, data scientist and AI system depends on.32 to 52 weeks74 lessons12 projects
- AIAI/ML EngineerTrain, ship and run machine learning and LLM systems that hold up in production.40 to 64 weeks63 lessons12 projects
- AIForward Deployed EngineerBuild AI systems inside customers' businesses and make them work in the real world.40 to 64 weeks49 lessons12 projects
- SoftwareSoftware Engineer (Backend)Build the services, APIs and data that products run on, and keep them fast and reliable.36 to 60 weeks69 lessons12 projects
- InfraDevOps / Platform EngineerBuild the pipelines, infrastructure and guard rails that let teams ship safely and often.36 to 60 weeks47 lessons12 projects
- InfraCloud and Network EngineerDesign, build and secure the networks and cloud foundations everything else runs on.36 to 60 weeks39 lessons12 projects
- SecurityCybersecurity EngineerDetect, respond to and prevent attacks, and help every team build securely.36 to 60 weeks35 lessons12 projects
- Web3Blockchain DeveloperWrite, test and ship smart contracts and the applications around them, where mistakes are public and permanent.40 to 64 weeks49 lessons12 projects
Data 16
- SQLLanguageAsk precise questions of tables of data. The one skill that appears in almost every data job description, and in many software and AI ones too.
- ETL and data pipelinesPracticeMove data from sources into an analysis-ready shape, reliably and repeatably. The core of data engineering and a growing part of analyst work.
- Apache SparkToolProcess data too big for one machine with Spark DataFrames, and keep jobs fast by understanding partitions, shuffles and joins.
- Data cleaning and qualityPracticeFind what is wrong with a dataset before it reaches a report, and write checks that catch it automatically next time.
- ExcelToolPivot tables, lookups and Power Query. Named in more than half of junior analyst postings, and still the fastest way to answer a one-off question.
- pandasToolThe Python library for tables. Load, clean, group, join and reshape data in a notebook or a script.
- StatisticsConceptDescribe data honestly, measure uncertainty, run fair experiments and fit regressions you can explain. The reasoning layer under every analyst and data scientist role.
- Data modellingConceptDesign tables that make analysis fast, correct and easy: facts, dimensions, grain and history.
- Data visualisation and storytellingPracticeChoose the right chart, design dashboards people actually use, and turn analysis into a decision. Mentioned in 94% of the Data Analyst job descriptions we read.
- Kafka and event streamingToolMove events between systems in real time with Kafka, and understand partitions, consumer groups and delivery guarantees.
- Apache Airflow and orchestrationToolSchedule, retry and monitor data pipelines as code, and run backfills safely. Taught with Airflow 3.
- Business acumenConceptUnderstand how a business makes money, which metrics matter, and how to turn a vague request into a question data can answer.
- Data warehousing and dbtConceptHow analytical warehouses store and scan data, how to design tables for cost and speed, and how to build tested transformations with dbt.
- Power BIToolBuild interactive reports on a proper data model, write DAX measures, and publish securely. Named in 80% of the Data Analyst job descriptions we analysed.
- RLanguageThe statistics-first language used in research, pharma, banking risk and many data science teams. Read, wrangle, model and plot with the tidyverse.
- TableauToolBuild interactive visual analysis and dashboards, and publish them free on Tableau Public for your portfolio.
AI 9
- Machine learningConceptTrain, evaluate and ship predictive models that are honest about how well they work. Named in 84% of Data Scientist and every AI/ML Engineer job description we read.
- LLMs and NLPConceptWork with text, from classic NLP to large language models. Call LLM APIs reliably, get structured output, and choose between prompting, retrieval and fine-tuning.
- AI agents and tool useConceptLet a model call your tools in a loop to finish multi-step tasks, and design the guard rails, limits and checks that make it safe to deploy.
- Deep learningConceptBuild and train neural networks in PyTorch, and know when they beat simpler models. Named in 68% of AI/ML Engineer postings, with PyTorch in 72%.
- LLM evaluationPracticeMeasure whether an LLM feature works, with labelled test sets, automatic checks and calibrated model graders, and stop regressions before release.
- MLOpsPracticeShip models as versioned, tested services, then watch them in production and retrain before they quietly fail.
- Model Context Protocol (MCP)ToolExpose your data and actions to any AI client through one open protocol, and run MCP servers securely.
- Prompt engineeringPracticeWrite clear, structured, versioned prompts that produce reliable behaviour, and improve them with evidence rather than guesswork.
- Retrieval-augmented generation (RAG)ConceptGround language model answers in your own documents, with citations, and measure whether retrieval actually finds the right text.
Infra 8
- Cloud platforms (AWS, Azure, Google Cloud)PlatformThe building blocks every cloud shares, how to use them safely from code, and how to design for reliability and cost. Taught on AWS, mapped to Azure and Google Cloud.
- CI/CDPracticeTest every change automatically and ship it safely, with pipelines as code. Taught with GitHub Actions; the ideas carry to GitLab CI, Jenkins and Azure Pipelines.
- Docker and containersToolPackage applications and models into small, secure container images, and run multi-service stacks locally with Compose.
- Monitoring and observabilityPracticeKnow what your systems are doing from their metrics, logs and traces, define service level objectives, and alert only on what matters.
- KubernetesPlatformRun containers reliably across machines: deployments, services, probes, resources, rollouts, autoscaling and day-two operations.
- Infrastructure as codePracticeDefine cloud infrastructure and server configuration in reviewed, versioned code with Terraform and Ansible, and run it safely across environments.
- Linux and BashPlatformThe operating system almost every server runs, and the shell you automate it with: files and permissions, processes and ports, logs, and safe Bash scripts.
- Networking fundamentalsConceptIP addressing and subnets, DNS, TCP and TLS, and the cloud networking built on them: VPCs, routing, NAT, security groups and load balancers.
Security 10
- Security compliance and data protectionPracticeUnderstand the frameworks and laws that shape security work in India and globally, protect personal data in pipelines, and turn controls into automated evidence.
- Security fundamentalsConceptThe ideas every engineer needs to build and run systems safely: confidentiality, integrity and availability, password hashing, encryption, secrets and defence in depth.
- Identity and access managementConceptDecide who can do what to which resource: least-privilege policies, roles with temporary credentials, explicit denies, and identities for machines as well as people.
- Applied cryptographyConceptUse cryptography correctly without inventing it: hashing, symmetric and public-key encryption, signatures, certificates and key management, and the mistakes to avoid.
- Application securityPracticeFind and fix the vulnerability classes in the OWASP Top 10 in your own code: injection, broken access control and weak cryptography, with static analysis in the pipeline.
- Incident responsePracticeRespond to a security incident in order: scope it with a timeline, contain it without destroying evidence, remove the attacker, recover, and learn.
- Penetration testingPracticeTest systems the way an attacker would, but with permission and a method: scope and authorise, recon, find and confirm weaknesses, and report them so they get fixed.
- SIEM and security operationsPracticeCollect and read security logs, triage alerts, and write detections as code that catch real attacks without drowning analysts in noise.
- Threat modellingPracticeFind security design flaws before they are built: diagram the system, ask what can go wrong with a method like STRIDE, and decide how to handle each threat.
- Vulnerability managementPracticeFind known vulnerabilities in dependencies, images and hosts, prioritise them by real-world risk rather than raw counts, and drive fixes to done.
Software 13
- PythonLanguageThe most requested language across data, AI, automation and backend roles. From your first script to code a team can maintain.
- Git and GitHubToolTrack changes, collaborate through branches and pull requests, and recover from mistakes without fear.
- REST APIs and HTTPConceptUnderstand HTTP well enough to debug any integration, and design APIs that other teams find predictable and safe.
- NoSQL databasesConceptDocument stores and key-value caches: when to use them instead of a relational database, how to model data for your queries, and how to index, cache and rate-limit.
- System designConceptDesign systems that meet their requirements at the scale they need: estimate load, choose building blocks, and explain the trade-offs out loud.
- TypeScript and JavaScriptLanguageThe language of the web and of many integration and SDK codebases. Typed JavaScript on Node.js: types, async code, validated API calls and tested services.
- Automated testingPracticeWrite fast unit tests and focused integration tests that catch real bugs, use test doubles for slow dependencies, and read coverage for what it does and does not tell you.
- GoLanguageA small, fast language for services and infrastructure tools, with concurrency built in. Much of the cloud-native tooling, including Kubernetes and Terraform, is written in it.
- Java and Spring BootLanguageThe most requested backend language in the Software Engineer postings we read, and the Spring Boot framework most of those teams use.
- Data structures and algorithmsConceptChoose the right data structure, reason about time and space complexity, and recognise the handful of patterns behind most coding interview problems.
- MicroservicesConceptSplit a system into services only where it helps, connect them with APIs and events, and keep them reliable when the network and other services fail.
- ReactToolBuild accessible, tested user interfaces with React components, state and props, with the patterns used in dApp and product front ends.
- RustLanguageA fast, memory-safe systems language used for blockchain clients, Solana programs, infrastructure tools and performance-critical services.
Web3 3
- Blockchain fundamentalsConceptHow blockchains actually work: hash-linked blocks, consensus, keys, addresses and signed transactions, built up from code you can run.
- DeFiConceptHow decentralised finance works under the hood: automated market makers, liquidity, slippage and price impact, lending and oracles, and the risks that come with them.
- Smart contracts and SolidityLanguageWrite, test, deploy and secure smart contracts in Solidity on the EVM, with Foundry for unit and fuzz tests, local chains and scripted interaction.
Working skills 4
- CommunicationWorking skillWrite and speak so busy people understand you the first time. Named in more job descriptions than any single tool across the roles we studied.
- Stakeholder managementWorking skillGather requirements, agree scope, manage expectations and keep people informed, so work lands with the people who asked for it.
- Agile ways of workingPracticeHow most engineering teams plan and ship work: sprints, boards, stand-ups and retrospectives, and how to be a good teammate inside them.
- Customer-facing deliveryPracticeRun discovery with customers, scope pilots with measurable success criteria, demo on their data and hand over systems they can run. The core of forward deployed work.
Data 28
- Beginner10 hNot startedAir quality sensor data explorationClean a year of hourly readings from a real city air quality station, where missing values are coded as -200.Data cleaning and qualityData visualisation and storytellingpandasPythonReal data · CC BY 4.0
- Beginner8 hNot startedBank marketing campaign performance reportAnalyse 45,211 real telemarketing calls and tell a bank which customers and channels to prioritise next time.CommunicationExcelStatisticsReal data · CC BY 4.0
- Beginner10 hNot startedCredit card default risk baselineExplore 30,000 real credit card customers and build an honest first model of who will default next month.Machine learningpandasPythonStatisticsReal data · CC BY 4.0
- Beginner8 hNot startedFrom a vague ask to a pilot planPlay the forward deployed engineer for an online store's support team, analyse their message backlog, and write the discovery notes and a one-page pilot plan with a measured baseline.CommunicationCustomer-facing deliverySQLStakeholder managementReal data · CDLA-Sharing-1.0
- Beginner10 hNot startedIndia development indicators on Tableau PublicPull real World Bank indicators through the API and publish an interactive dashboard comparing India with its peers.Data visualisation and storytellingPythonTableauReal data · CC BY 4.0
- Beginner8 hNot startedMonthly data quality report for taxi tripsBuild a repeatable data quality report for any month of NYC taxi data, with thresholds and a clear verdict.Data cleaning and qualityPythonSQLReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Beginner8 hNot startedNYC taxi demand in SQLAnswer ten real operations questions about 3.4 million New York taxi trips with clean, reviewed SQL.Data cleaning and qualitySQLReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Beginner10 hNot startedRetail sales KPI dashboardTurn a year of real online retail transactions into a one-page KPI dashboard a sales head can read in 30 seconds.Data cleaning and qualityData visualisation and storytellingExcelPower BIReal data · CC BY 4.0
- Beginner10 hNot startedRetail star schema in SQLModel two years of real online retail orders as a star schema with declared grains and a reconciliation query that proves it.Data cleaning and qualityData modellingSQLReal data · CC BY 4.0
- Intermediate16 hNot startedA year of taxi trips in SparkProcess twelve months of NYC taxi data with PySpark, write partitioned output, and document every performance decision.Data warehousing and dbtPythonApache SparkReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Intermediate12 hNot startedCustomer cohort retention analysisBuild monthly cohort retention tables in SQL from raw invoices and explain what they mean for a wholesaler.Business acumenData visualisation and storytellingSQLReal data · CC BY 4.0
- Intermediate16 hNot startedE-commerce star schema and DAX dashboardModel a real Brazilian marketplace's orders as a star schema and build a Power BI report with time intelligence.Data modellingPower BISQLReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Intermediate18 hNot startedMarketplace warehouse with historyTurn nine raw marketplace tables into a tested warehouse with order and item facts, and sellers that keep their history.Data cleaning and qualityData modellingData warehousing and dbtSQLReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Intermediate14 hNot startedOrchestrated monthly pipeline with AirflowSchedule, validate and backfill a year of taxi data with Airflow 3, with retries for late files and alerts on failure.Apache Airflow and orchestrationData cleaning and qualityETL and data pipelinesPythonReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Intermediate14 hNot startedPredict purchases from browsing sessionsModel 12,330 real e-commerce sessions to find which behaviour signals a purchase, and what the business can do with it.Data visualisation and storytellingMachine learningpandasPythonReal data · CC BY 4.0
- Intermediate20 hNot startedReliable order events with Kafka and the outbox patternPublish order events from a service to Kafka through a transactional outbox, consume them in two services idempotently, and prove nothing is lost or double-counted.Kafka and event streamingMicroservicesSQLAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Intermediate12 hNot startedRFM customer segmentation and win-back planSegment 4,339 real customers by recency, frequency and value in pandas, and write the win-back plan marketing will run.CommunicationpandasPythonReal data · CC BY 4.0
- Intermediate16 hNot startedTaxi analytics marts with dbtBuild a tested dbt project over a year of NYC taxi data, with staging, incremental facts, marts and documentation.Data modellingData warehousing and dbtGit and GitHubSQLReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Intermediate14 hNot startedTerm deposit propensity model for a call centreRank a bank's customers by likelihood to subscribe, choose a threshold from call centre capacity, and prove the lift.CommunicationMachine learningPythonStatisticsReal data · CC BY 4.0
- Intermediate12 hNot startedWhat drives bike-share demand?Explain two years of real Capital Bikeshare demand with regression, and turn it into guidance for an operations team.Data visualisation and storytellingPythonStatisticsReal data · CC BY 4.0
- Advanced24 hNot startedA small cloud data platform on a budgetLand, catalogue and query real data in the cloud with infrastructure as code, least-privilege access and a hard budget.Cloud platforms (AWS, Azure, Google Cloud)Data warehousing and dbtETL and data pipelinesInfrastructure as codeReal data · NYC Open Data public data (TLC terms of use, no warranty), CC BY 4.0
- Advanced24 hNot startedAutomated taxi operations pipeline and dashboardBuild a monthly pipeline over NYC taxi data with data tests, and a dashboard that refreshes itself.Data cleaning and qualityETL and data pipelinesPower BISQLReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Advanced20 hNot startedExperiment readout on 14 million real ad impressionsAnalyse a real randomised advertising experiment, check it is trustworthy, and write the readout a product lead signs off.CommunicationPythonStatisticsReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Advanced20 hNot startedHourly bike demand forecasting with backtestingForecast next-day hourly rentals for a real bike-share system, and prove the forecast with a proper backtest.CommunicationMachine learningPythonStatisticsReal data · CC BY 4.0
- Advanced20 hNot startedIncremental warehouse loads with late and corrected dataSimulate daily order drops with late and corrected rows, and build incremental loads that stay correct on every rerun.Data modellingData warehousing and dbtETL and data pipelinesSQLReal data · CC BY 4.0
- Advanced24 hNot startedReal-time taxi demand streamReplay real taxi trips through Kafka, compute live demand per zone, and prove the pipeline survives crashes without losing or double-counting events.ETL and data pipelinesKafka and event streamingMonitoring and observabilityPythonReal data · NYC Open Data public data (TLC terms of use, no warranty)
- Advanced20 hNot startedSecure self-serve sales model for 27 regional managersTurn a report into a governed, documented semantic model with row-level security, refresh and a release process.Data modellingPower BIStakeholder managementReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Advanced20 hNot startedWhat developers in India earn and whyClean the real Stack Overflow Developer Survey, compare pay for Indian and global respondents, and publish a careful, sourced article.CommunicationData visualisation and storytellingpandasStatisticsReal data · ODbL 1.0 (database), DbCL 1.0 (contents)
AI 11
- Beginner10 hNot startedA support reply assistant with structured outputUse an LLM to classify customer messages and draft replies with structured output, then measure it on 100 labelled messages before anyone sees a draft.LLMs and NLPPrompt engineeringPythonReal data · CDLA-Sharing-1.0
- Beginner10 hNot startedRoute banking questions to 77 intentsClassify 13,083 real customer banking queries into 77 intents, compare TF-IDF with sentence embeddings, and study the mistakes that remain.LLMs and NLPMachine learningPythonReal data · CC BY 4.0
- Beginner8 hNot startedSMS spam classifier with an honest baselineBuild and compare text classifiers on 5,572 real SMS messages, and explain why the simple one wins.LLMs and NLPMachine learningPythonReal data · CC BY 4.0
- Intermediate20 hNot startedA cited question-answering assistant over the Kubernetes docsBuild retrieval over the official Kubernetes documentation, answer questions with citations, and measure retrieval and answer quality on your own test set.LLMs and NLPLLM evaluationPythonRetrieval-augmented generation (RAG)Real data · CC BY 4.0
- Intermediate16 hNot startedAn evaluation harness that blocks bad prompt changesTurn an LLM ticket-triage prompt into a tested component, with a golden set, code-based graders, a calibrated LLM judge and a CI gate that fails on regressions.CI/CDLLMs and NLPLLM evaluationPrompt engineering+1Real data · CC BY 4.0
- Intermediate14 hNot startedAn MCP server for World Bank development dataExpose the World Bank Indicators API to AI assistants through a small, well-described MCP server with tests, limits and clear errors.AI agents and tool usePythonModel Context Protocol (MCP)REST APIs and HTTPReal data · CC BY 4.0
- Intermediate16 hNot startedLand cover from satellite images with transfer learningClassify 27,000 Sentinel-2 image patches into 10 land cover classes, first with a small CNN from scratch and then by fine-tuning a pretrained network.Deep learningMachine learningPythonReal data · MIT (Copernicus Sentinel data terms also apply)
- Advanced28 hNot startedA data question agent with guardrails and an eval suiteBuild a tool-using agent that answers questions about NYC taxi trips with read-only SQL, cites the query it ran, stays within budgets and is measured on a task suite.AI agents and tool useLLM evaluationMonitoring and observabilityPython+1Real data · NYC Open Data public data (TLC terms of use, no warranty)
- Advanced24 hNot startedFine-tuning versus prompting, decided with numbersSolve the same 77-intent task three ways (prompting an LLM, embeddings plus a classifier, fine-tuning a small transformer) and recommend one on accuracy, cost, latency and maintenance.Deep learningLLMs and NLPLLM evaluationMachine learningReal data · CC BY 4.0
- Advanced32 hNot startedRun a customer pilot end to endTake a simulated customer from discovery to a decision: scope a pilot, build a cited support assistant over their policies, evaluate it, demo it on their data and hand it over.Customer-facing deliveryLLMs and NLPLLM evaluationMonitoring and observability+2Real data · CDLA-Sharing-1.0
- Advanced24 hNot startedShip a model and catch it failingServe a propensity model as a tested API, simulate two years of traffic, detect drift and write the retraining policy.Cloud platforms (AWS, Azure, Google Cloud)Machine learningMLOpsPythonReal data · CC BY 4.0
Infra 25
- Beginner12 hNot startedA production-like Compose stack with a TLS reverse proxyRun a web app, its database and a reverse proxy with TLS in Docker Compose, with health checks, restart policies, private networks and persistent volumes.Linux and BashNetworking fundamentalsDocker and containersReal data · Open Database License (ODbL); images CC BY-SA
- Beginner10 hNot startedA vulnerability management pipelineScan an application's dependencies and container image, prioritise findings by real risk, and produce a tracked, deadline-driven remediation plan.CI/CDPythonVulnerability management
- Beginner8 hNot startedAn IP address plan for a growing companyDesign a non-overlapping address plan for offices, cloud accounts and Kubernetes clusters, and write a checker that catches overlaps and exhausted ranges.Networking fundamentalsPython
- Beginner10 hNot startedAn operations toolkit in Bash with testsWrite the small scripts an on-call engineer needs (health checks, log clean-up, disk alerts, a deploy helper), make them safe and test them in CI.CI/CDGit and GitHubLinux and BashReal data · Open Database License (ODbL); images CC BY-SA
- Beginner10 hNot startedContainerise a service and ship it through CIPackage the orders API in a small, secure container image, run it with its database in Docker Compose, and build, test and scan it on every pull request.CI/CDGit and GitHubDocker and containersAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Beginner12 hNot startedHarden a Linux server and prove your backups restoreTake a fresh Linux VM running a small catalogue app to a hardened, monitored, backed-up state, and restore it from scratch to prove the backups work.Linux and BashNetworking fundamentalsSecurity fundamentalsReal data · Open Database License (ODbL); images CC BY-SA
- Beginner8 hNot startedHost a static site with a CDN, TLS and your own domainServe a static site from object storage through a CDN with HTTPS and DNS, with the bucket private and deployments done by a least-privilege CI role.Cloud platforms (AWS, Azure, Google Cloud)Identity and access managementNetworking fundamentals
- Beginner12 hNot startedTriage a day of logs and write detectionsAnalyse a set of security logs, separate noise from a real intrusion, and write and test Sigma detections for what you find.Linux and BashPythonSIEM and security operations
- Intermediate20 hNot startedA cloud network and service in Terraform, reviewed like codeDefine a two-zone VPC, a load-balanced container service and its IAM roles in Terraform modules, with remote state, plans in pull requests and policy checks.Cloud platforms (AWS, Azure, Google Cloud)Identity and access managementNetworking fundamentalsInfrastructure as codeReal data · Open Database License (ODbL); images CC BY-SA
- Intermediate14 hNot startedA site-to-site VPN and private DNS labConnect two simulated sites with a WireGuard tunnel, route between their subnets, and make private names resolve on both sides.Linux and BashNetworking fundamentals
- Intermediate20 hNot startedAn LLM triage integration driven by helpdesk webhooksBuild a TypeScript service that receives signed helpdesk webhooks, triages each ticket with an LLM, writes the result back through the API, and survives retries, duplicates and outages.LLMs and NLPMonitoring and observabilityDocker and containersREST APIs and HTTP+1Real data · CDLA-Sharing-1.0
- Intermediate14 hNot startedConfigure a fleet of servers idempotently with AnsibleWrite Ansible roles that configure web and database servers from scratch, run them twice with no changes, and test them with Molecule.Linux and BashSecurity fundamentalsInfrastructure as codeReal data · Open Database License (ODbL); images CC BY-SA
- Intermediate20 hNot startedGitOps delivery to Kubernetes with Helm and Argo CDPackage a service as a Helm chart, deploy it to dev and prod namespaces through Argo CD, and promote releases with pull requests instead of kubectl.CI/CDGit and GitHubDocker and containersKubernetesReal data · Open Database License (ODbL); images CC BY-SA
- Intermediate16 hNot startedMonitoring, SLOs and alerts that wake the right personInstrument a service, collect metrics and logs, define SLOs, build multi-window burn-rate alerts and a dashboard, and test the alerts with injected failures.Linux and BashMonitoring and observabilityKubernetesReal data · Open Database License (ODbL); images CC BY-SA
- Intermediate16 hNot startedZero-trust networking for a microservices app on KubernetesDeploy an 11-service demo shop on a local cluster, apply default-deny network policies, allow only the flows the app needs, and prove unwanted connections fail.Networking fundamentalsKubernetesSecurity fundamentalsReal data · Apache License 2.0
- Advanced16 hNot startedA cloud cost review with savings you can proveEstimate and measure the cost of running a demo shop, find the biggest savings, implement three of them, and present the trade-offs to a budget owner.Cloud platforms (AWS, Azure, Google Cloud)Monitoring and observabilityStakeholder managementReal data · Apache License 2.0
- Advanced24 hNot startedA golden path: a service template with everything built inBuild a template that gives a new service CI, infrastructure, deployment, monitoring and documentation in minutes, and measure how long it takes a developer to ship with it.CI/CDMonitoring and observabilityKubernetesStakeholder management+1Real data · Open Database License (ODbL); images CC BY-SA
- Advanced24 hNot startedA hub-and-spoke cloud network in TerraformDesign and code a transit hub connecting shared services, prod and dev VPCs and an office VPN, with segmented route tables, central egress, private DNS and flow logs.Cloud platforms (AWS, Azure, Google Cloud)Networking fundamentalsMonitoring and observabilityInfrastructure as code
- Advanced32 hNot startedE-commerce microservices on Kubernetes that survive failureSplit a shop into orders, catalogue and payments services, run them on a local Kubernetes cluster, and show that one failing service does not take down the rest.MicroservicesMonitoring and observabilityDocker and containersKubernetes+1Real data · CC BY-NC-SA 4.0 (non-commercial use, share alike), Open Database License (ODbL); images CC BY-SA
- Advanced24 hNot startedHigh availability and disaster recovery you have testedRun a demo shop across availability zones, define recovery objectives, and prove them by failing a zone and restoring from backups in another region.Cloud platforms (AWS, Azure, Google Cloud)Monitoring and observabilityKubernetesInfrastructure as codeReal data · Apache License 2.0
- Advanced20 hNot startedOptimise and harden your own contractsTake the contracts you built earlier, measure gas, run static analysis and coverage in CI, and ship an improved version with every change justified by numbers.CI/CDSmart contracts and SolidityAutomated testingReal data · AGPL-3.0
- Advanced24 hNot startedPass a customer's security review with an LLM deploymentDeploy an LLM support assistant into a customer-controlled cloud account with least-privilege access, PII redaction, audit logs and data retention, and answer the security questionnaire with evidence.Cloud platforms (AWS, Azure, Google Cloud)Security compliance and data protectionCustomer-facing deliveryLLMs and NLP+1Real data · CDLA-Sharing-1.0
- Advanced16 hNot startedRun a game day: break things, respond, write the postmortemPlan and run failure experiments against a small production-like system, respond as an on-call team, and write blameless postmortems with actions that are then implemented.CommunicationLinux and BashMonitoring and observabilityKubernetesReal data · Open Database License (ODbL); images CC BY-SA
- Advanced24 hNot startedSafe deploys with canaries and automatic rollbackShip a bad release on purpose and show that a canary with automated analysis stops it before most users see it, then roll back without a human.CI/CDMonitoring and observabilityKubernetesReal data · Open Database License (ODbL); images CC BY-SA
- Advanced24 hNot startedServe a model on Kubernetes with autoscaling and a safe rolloutPackage an intent classifier as a container, run it on a local Kubernetes cluster with health checks, autoscaling and metrics, and roll out a new version without errors.CI/CDMLOpsMonitoring and observabilityDocker and containers+1Real data · CC BY 4.0
Security 13
- Intermediate18 hNot startedA multi-signature walletBuild a wallet that needs M of N owners to approve a transaction, with on-chain confirmations or EIP-712 signatures, replay protection and full tests.Applied cryptographySmart contracts and SolidityAutomated testingReal data · MIT
- Intermediate14 hNot startedAudit and tighten cloud permissionsFind over-privileged users and roles in a sample cloud account, replace them with scoped policies, and prove with tests that the workloads still work and nothing more is allowed.Security compliance and data protectionIdentity and access managementPythonReal data · Apache License 2.0
- Intermediate18 hNot startedGuardrails that make insecure cloud setups impossibleBuild preventative and detective controls for a cloud account so common mistakes are blocked or caught automatically, and prove they work.Security compliance and data protectionIdentity and access managementSecurity fundamentalsInfrastructure as code
- Intermediate14 hNot startedHarden a host and measure it against a benchmarkHarden a Linux server against a CIS-style benchmark, automate the changes, and produce before-and-after compliance scores with evidence.Security compliance and data protectionLinux and BashSecurity fundamentals
- Intermediate16 hNot startedRun an incident from alert to reportGiven an intrusion in a set of logs and artifacts, scope it, build a full timeline, decide containment and recovery, and write a blameless postmortem.CommunicationIncident responseSIEM and security operations
- Advanced20 hNot startedA cloud security posture review with fixes as codeScan infrastructure code and a sandbox account against CIS benchmarks, prioritise findings, fix them in Terraform and add checks to CI so they cannot come back.Security compliance and data protectionIdentity and access managementSecurity fundamentalsInfrastructure as codeReal data · Apache License 2.0
- Advanced26 hNot startedA detection engineering lab with tested coverageStand up a SIEM, ingest realistic telemetry, write a library of tested detections mapped to ATT&CK, and measure your coverage and false-positive rate.Incident responseMonitoring and observabilityPythonSIEM and security operations
- Advanced22 hNot startedA field-encryption and secrets service done rightBuild a small service that encrypts sensitive fields with envelope encryption, manages keys with rotation, and stores application secrets safely, with tests for the failure cases.Applied cryptographyIdentity and access managementSecurity fundamentalsAutomated testing
- Advanced28 hNot startedA full security assessment of a practice web appThreat model, test and report on OWASP Juice Shop end to end, then write the code fixes and detections a defender would add.Application securityCommunicationPenetration testingThreat modellingReal data · MIT
- Advanced26 hNot startedA purple-team exercise: attack, detect, improveRun a coordinated exercise where authorised attack techniques are executed on a lab, measured against your detections, and the gaps are closed and retested.CommunicationIncident responsePenetration testingSIEM and security operations
- Advanced24 hNot startedA secure software delivery pipelineAdd security gates across a delivery pipeline (secrets, SAST, dependencies, container, IaC and signing) so insecure code and artifacts cannot reach production, with an exception process.Application securityCI/CDSecurity compliance and data protectionIdentity and access management
- Advanced20 hNot startedA secure software supply chain with policy as codeSign images, generate SBOMs, scan dependencies, authenticate CI with OpenID Connect and enforce policies at build and deploy time so only trusted artifacts reach the cluster.CI/CDSecurity compliance and data protectionIdentity and access managementKubernetes+1Real data · Open Database License (ODbL); images CC BY-SA
- Advanced16 hNot startedDesign and pilot a security champion programDesign a program that spreads security ownership into engineering teams, pilot it with one team, and measure whether it changed behaviour.Application securityCommunicationStakeholder managementThreat modelling
Software 16
- Beginner14 hNot startedA dApp front end for the tip jarBuild a React front end that connects a wallet, reads contract state, sends a transaction and shows every state from pending to confirmed or failed.ReactSmart contracts and SolidityTypeScript and JavaScript
- Beginner14 hNot startedA product catalogue API in GoServe a searchable catalogue of real food products from Open Food Facts with Go's standard library, table-driven tests and graceful shutdown.GoREST APIs and HTTPAutomated testingReal data · Open Database License (ODbL); images CC BY-SA
- Beginner14 hNot startedA Rust command-line tool for chain dataWrite a Rust CLI that queries an Ethereum JSON-RPC node for blocks and balances, parses the responses safely, and is covered by tests against recorded responses.REST APIs and HTTPRustAutomated testing
- Beginner16 hNot startedAn orders REST API with Spring Boot and PostgreSQLLoad real e-commerce orders into PostgreSQL and serve them through a validated, paginated, tested Spring Boot API.Java and Spring BootREST APIs and HTTPSQLAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Beginner10 hNot startedAPI to data lake ingestionPull development indicators from the World Bank API into a raw and clean lake layout, with retries, pagination and logging.ETL and data pipelinesPythonREST APIs and HTTPReal data · CC BY 4.0
- Beginner12 hNot startedBuild a toy blockchain with Merkle proofsExtend a hash-linked chain with proof of work, Merkle roots, transaction signatures and a validity checker, and prove each property with tests.Blockchain fundamentalsPythonAutomated testing
- Beginner12 hNot startedFind and fix vulnerabilities in a small appReview a deliberately insecure web app, confirm each flaw with a safe test on your own instance, fix it, and prove the fix with an automated test.Application securityPythonAutomated testing
- Beginner10 hNot startedPut a safety net under legacy codeWrite characterization tests for an untested order-pricing module using real order data, then refactor it without changing behaviour.Git and GitHubPythonAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Beginner12 hNot startedSync support tickets into a helpdesk API with TypeScriptWrite a typed, validated and tested TypeScript tool that loads customer messages into a helpdesk REST API without duplicates, even when the API fails halfway.Git and GitHubREST APIs and HTTPTypeScript and JavaScriptReal data · CDLA-Sharing-1.0
- Intermediate16 hNot startedA fast catalogue with MongoDB and a Redis cacheModel food products as documents, index them for real queries, add a cache in front, and prove the speed-up with a load test and numbers.NoSQL databasesMonitoring and observabilityREST APIs and HTTPAutomated testingReal data · Open Database License (ODbL); images CC BY-SA
- Intermediate14 hNot startedFind and fix a slow endpoint with measurementsLoad test an API over real orders, trace where the time goes, fix the bottlenecks one at a time, and show each improvement with numbers.Monitoring and observabilitySQLSystem designAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Intermediate14 hNot startedSecure an API with tokens, roles and rate limitsAdd login, short-lived access tokens, role-based access and per-user rate limits to a catalogue API, and test every rule.NoSQL databasesREST APIs and HTTPAutomated testingReal data · Open Database License (ODbL); images CC BY-SA
- Advanced28 hNot startedA full dApp with an event indexer and APIShip a complete application: contracts on a testnet, an indexer that turns events into a database, an API, and a React front end, with handling for reorganisations.ReactSmart contracts and SoliditySystem designTypeScript and JavaScript
- Advanced28 hNot startedBuild the notification service from the designImplement a notification service with idempotent requests, per-priority queues, provider rate limits, retries, fallbacks and a dead-letter queue, driven by real order events.Kafka and event streamingMicroservicesMonitoring and observabilitySystem designReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Advanced16 hNot startedChange a live database schema with zero downtimeSplit a column in a busy orders table using the expand and contract pattern, with a backfill, dual writes and a rollback plan, while a load test keeps running.CI/CDJava and Spring BootSQLAutomated testingReal data · CC BY-NC-SA 4.0 (non-commercial use, share alike)
- Advanced30 hNot startedDesign and build catalogue search for millions of productsWrite a system design for product search over the full Open Food Facts catalogue, implement the critical path, and load test it against the targets in your design.GoNoSQL databasesMonitoring and observabilitySystem designReal data · Open Database License (ODbL); images CC BY-SA
Web3 6
- Beginner12 hNot startedAn ERC-20 token with a full test suiteBuild an ERC-20 token on OpenZeppelin, add capped minting and pausing with access control, and cover it with unit, fuzz and invariant tests in Foundry.Blockchain fundamentalsSmart contracts and SolidityAutomated testingReal data · MIT
- Intermediate22 hNot startedA constant-product AMM in SolidityImplement a minimal constant-product pool with liquidity shares, swaps with fees and slippage limits, and test it against a Python reference model.DeFiSmart contracts and SolidityAutomated testingReal data · GPL-3.0
- Intermediate18 hNot startedAn escrow contract with disputes and deadlinesBuild a two-party escrow with an arbiter, deadlines and refunds, model its states explicitly, and prove its safety properties with invariant tests.DeFiSmart contracts and SolidityAutomated testingReal data · MIT
- Intermediate16 hNot startedAn NFT collection with an allowlist mintBuild an ERC-721 collection with a Merkle-proof allowlist, a public mint phase, royalties metadata and withdrawals, fully tested in Foundry.Blockchain fundamentalsSmart contracts and SolidityAutomated testingReal data · MIT
- Advanced24 hNot startedA staking rewards contract with fair accountingBuild a contract that pays rewards to stakers pro rata over time, using the reward-per-token accumulator pattern, and prove the accounting with invariant tests.DeFiSmart contracts and SolidityAutomated testingReal data · MIT
- Advanced24 hNot startedOn-chain governance with proposals, voting and a time-lockBuild a governance system with a voting token, proposals, quorum, a time-lock and execution, and test the full lifecycle of a proposal.Blockchain fundamentalsSmart contracts and SolidityStakeholder managementReal data · MIT