Start here
A placement preparation assistant
Track applications, understand job requirements, and turn preparation notes into a weekly plan.
See how you'll build itNew cohortstarts Sat, 17 OctAI Engineer CohortBook your seat · ₹499
Learn to buildwith AI,not around it.
SkillCaptain is where students learn to become strong engineers. You learn the fundamentals and the AI tools engineers use today by building real projects, with a mentor guiding you in Hindi or English.
Agent session~/study-planner
↵
src/TaskList.tsxdiff
Previewlocalhost:5173/tasks
This week3 tasks
- OS lab recorddue in 6 days
- Mock interview prepdue in 4 days
- DBMS assignmentdue in 2 days
How learning works here: you build with AI, then a mentor makes sure you understand what you built. Tap a project to see another lesson.
Modern tools.
Real skills.
- Codex
- Claude Code
- OpenCode
- MCP
- n8n
- GitHub
Start with a problem.
Finish with something useful.
Pick one project and grow it through the whole path. Every new skill adds something you can use.
For the curious learner
A study companion with sources
Ask questions about your notes. Get grounded answers, see the evidence, and spot what's missing.
Explore document searchFor the repeatable task
A weekly planning workflow
Bring updates together, draft an action plan, and keep a person in control before anything is shared.
Explore the workflowA little code.
A lot of curiosity.
This path is for you if you can read a small program, edit a function, and run a project locally. AI concepts arrive exactly when your project needs them.
- No prior ML experience
- No advanced maths
- 1–2 hrs practice between builds
12 builds.
Something to show for each one.
Learn the workflow. Build it yourself. Check that it works. Repeat.
Builds 01–04
Build with AI
01PromptingGet better answers from AI
Give the model a clear task, useful context, and a definition of a good result.
- Task, context and constraints
- Examples and structured instructions
- Comparing results against a clear brief
You leave with: A tested prompt set and a one-page brief for your app.
02Codex · Claude Code · OpenCodeBuild your first app with a coding agent
Watch the same task in three tools. Pick one for your hands-on build.
- Codex: plan, build and run
- Claude Code: inspect edits and refine
- OpenCode: explore and verify
- Hands-on: build with your chosen agent
You leave with: A working application tracker and a tool choice you can explain.
03AI pair programmingFix the bug. Understand the fix.
Give the agent a reproducible problem, review its changes, and test the result.
- Reproduction steps and error messages
- Reviewing a diff
- Tests that catch the original failure
You leave with: A verified bug fix, a regression test, and a reviewed change.
04Context engineeringMake AI understand your project
Keep the project rules and relevant context available to a fresh agent session.
- AGENTS.md and CLAUDE.md
- Relevant files and acceptance criteria
- A fresh-session check
You leave with: Project instructions that improve the next change.
Builds 05–08
Give it real capabilities
05Agent skillsTeach your assistant a repeatable workflow
Turn a useful review procedure into a skill you can reuse.
- When the skill should run
- Inputs, steps and output format
- Testing the boundaries
You leave with: A reusable skill checked against normal and incomplete inputs.
06AI pluginsAdd a new capability with a plugin
Extend your chosen assistant and use the addition on a real project task.
- Checking compatibility and permissions
- Installing and configuring a plugin
- Verifying its output
You leave with: An installed plugin that completes a useful piece of work.
07MCPConnect AI to your tools
Let an assistant use actual project records through a controlled tool connection.
- Clients, servers and tools in practice
- Read-only access first
- Handling a failed tool call
You leave with: A working MCP connection and a plan grounded in tool results.
08LLM APIs · Structured outputsPut an AI feature inside your app
Call a model from your backend and turn its response into validated application data.
- Server-side API calls
- Validating JSON output
- Timeouts, retries and secret handling
You leave with: A useful AI feature with input validation and error handling.
Builds 09–12
Test it. Then ship it.
09RAG · Document searchChat with your notes — with sources
Retrieve relevant passages before answering, and acknowledge missing information.
- Chunking and retrieval through an example
- Grounded answers and citations
- Questions your notes cannot answer
You leave with: A document assistant with source references and an unknown-answer path.
10AI agents · n8nTurn a recurring task into a workflow
Collect records, draft a weekly action plan, and add a human review step.
- A bounded agent task
- Tool inputs and outputs
- Approval and failure paths
You leave with: An automation you can inspect and approve before it acts.
11AI evals · Reliability · CostFind out whether your AI actually works
Write a small test set, inspect failures, and compare quality and cost after a change.
- Expected behaviour and scoring
- Error analysis
- Cost per task and tradeoffs
You leave with: An evaluation report with a measured improvement.
12Deployment · PortfolioDeploy it. Demo it. Make it yours.
Put the app online, test it with peers, and explain what you built.
- Deployment and a smoke test
- README and architecture notes
- A demo with honest limitations
You leave with: A deployed app, a source repository, feedback, and a portfolio-ready demo.
AI sounds confident.
Should you trust it?
Try a tiny piece of the path. Check whether an answer is supported by its source before you use it.
Interactive sampleNo signup needed
Your first AI evaluationBuild 11 preview
The source
“The workshop starts at 10 am. Bring a laptop.”
The question
When does it start, and is lunch included?
Pick an answer to see what the evidence supports.
Watch it. Build it.
Make it better.
Leave with work you understand — and enough confidence to keep building after the path.
- 01
See the whole workflow
Follow a real task, including the decisions, errors and fixes along the way.
- 02
Build your version
Use guided exercises and small checkpoints to add a useful capability to your project.
- 03
Show that it works
Review the output, run your checks, and use feedback to improve the next version.
Your portfolio, by demo day
- Deployed app
- Source repository
- Evaluation report
- Project walkthrough
Guidance when it matters.
Learn with working developers and full-time mentors. Bring the questions, the broken builds, and the “why did it do that?” moments.
1-on-1 mentorship
Personal guidance in Hindi or English — on your code, your project, and your next step.
5 tutor sessions a week
Private sessions every week with the Pro plan, so you never stay stuck for long.
Weekly project feedback
Regular reviews of what you built, with concrete fixes to make before the next build.
Priority help, on request
Doubts answered when you need them — on the concept, the tool, or the error message.
“When I started at SkillCaptain, I was a manual tester who had never written a line of code.”
Ekta KumariSkillCaptain learner
Watch story“SkillCaptain cleared the confusion I had and gave me a clear path to follow.”
Vedant SalviSkillCaptain learner
Watch story“When I joined SkillCaptain, it was like finally having someone in my corner.”
Arvind KumarSkillCaptain learner
Watch storyStart free. Go Pro when you're ready.
One plan, everything included — the 12-week AI builder path with weekly live reviews. 5 free classes to start — pay only after.
Save 17%
MOST POPULAR
SkillCaptain Pro
Everything you need to get job ready
₹1,999₹1,666/mo
Billed ₹4,999 once every 3 months · save 17%
- 5 private tutor sessions a week
- Full course and content access
- Masterclasses and webinars
- Weekly assessments
- Expert mock interview every month
- Full resume, ATS and AI interview tools
- A portfolio built for you
- Priority help, on request
5 free classes to start. Pay only after.
Starts with an orientation call, then 5 free trial classes with full access. No payment is collected on this page. See plans for billing, taxes and refund terms.
Who is this for?
Students and early-career developers who can read a small program, edit a function, and run a project locally. If you have never written code, start with introductory practice and talk to us about your starting point.
Do I need machine-learning experience or advanced maths?
No. Concepts arrive through the task you are building. The focus is using models, coding agents, tools, and evaluations — not training models from scratch.
Will I code with AI, or build AI into an app?
Both. First, use a coding agent to build and improve an application. Then add model APIs, document retrieval, tool connections, and an agent workflow. Finish by testing and deploying the result.
What are skills, plugins, and MCP?
A skill is a reusable procedure for an assistant. A plugin packages capabilities for a particular tool. Model Context Protocol (MCP) connects an AI application to external tools and data. You practise all three.
Do I need paid AI tools to start?
No. You begin on free tiers. Subscriptions, model API usage, and hosting are separate costs — check the supported setup and any extra budget before you scale up.
What will I have at the end?
The target is one deployed AI app, its source repository, a reusable skill, a working tool integration, an evaluation report, and a short demo. You need to complete the builds to reach those outcomes.
How much time should I set aside?
The path runs about 12 weeks at roughly one build a week, with a weekly live review and a demo day at the end. Plan for the builds plus 1–2 hours of practice between them.
What does it cost?
You start with an orientation call, then 5 free trial classes with full access. The AI builder path is included in the Pro plan — see the plans section for billing periods, taxes, and what's included.
What if I need a refund?
Refunds follow our published cancellation and refund policy. Read it before paying, and talk to us if anything is unclear.
Does this guarantee a job?
No. The outcome is practical project experience you can explain and demonstrate. Hiring depends on your broader skills, experience, and the employer's requirements.