Milestone 1
A tested, documented CLI tool that pulls from a real API, stores results in Postgres, and ships with a README a stranger could follow.
Phase 0 · Weeks 1–4
# Flagship track
AI Engineering
We train software engineers who build products with AI, not researchers who build AI.
Weekly rhythm
01
Learn
Live-coded sessions, key concepts, architecture & tools.
MON – THU02
Build
Lab day — students build, instructors circulate.
FRI03
Review
Code review and demo. Every student presents something every week.
SAT04
Ship
Off. Burnout is a placement risk.
SUNAssumed load
20 hours per week
Roughly 8 hours of instruction, 10 hours of lab and project work, and 2 hours of review.
Cohort size
Cap at 30 students
One instructor and one reviewer. Code review is the product; it does not scale past this.
~/tekschool/ai-engineering
AI Product Architecture
Business Problem
Define the goal & user need
AI System
Orchestrate model + data + tools
Model
LLM / Embeddings
Data
Vector DB / APIs
Tools
Integrations
Testing & Evaluation
Quality • Safety • Performance
Production
Deploy • Monitor • Iterate
latency 412mstokens 284cost ₹0.021
A graduate can take a vague business requirement, design a system around a model they did not train, ship it to production, measure whether it works, and keep it running within a cost budget.
What's included
6 months instruction + 3 months embedded internship with a parallel learning track.
Cohort size: cap at 30 with one instructor and one reviewer. Code review is the product; it does not scale past this.
Six milestone projects across Phases 0–5, assessed by an engineer against a written rubric.
Phases 3–5: someone currently shipping AI in production. Non-negotiable.
Internship exit requirements: a shipped feature in production, a client-facing reference or documented performance rating, a written case study, and a portfolio of three defensible projects.
Intervention rule: a student failing two consecutive milestones gets a structured remediation plan, not a pass.
# Honest filter
Is this the right room for you?
We'd rather you self-select out now than three months in.
# Perfect if
- Freshers from any engineering or science background.
- No ML, maths or AI prerequisite.
- Required: ability to write and debug a program in any language, tested at admission.
- You want to build products people use, not publish research.
# Not for you if
- Training models from scratch, and deep learning mathematics.→ Take ML Engineering instead
- Classical ML algorithm derivations.
- Data science, statistics and analytics.→ Take Data Engineering instead
# Curriculum
How the 6 + 3 months of internship flows
Start
Week 1 — Python/MERN as an engineer writes it.
Finish
Weeks 25–36 — embedded internship: four days a week on client work, one day on the learning track.
Schedule Curriculum details Week 1 Python/MERN as an engineer writes it: environments and packaging (uv or poetry), modules and imports, type hints, dataclasses, error handling and custom exceptions, logging instead of print, virtual environments, project layout. Lab: restructure a messy single-file script into a proper package. Week 2 Testing, Git, and working in a team: pytest, fixtures, parametrisation, mocking external calls, test coverage as a signal not a target. Git branching, meaningful commits, pull requests, resolving conflicts, reading a diff. Week 3 The web, APIs and data: HTTP verbs, status codes, headers, idempotency. REST design. JSON and schema validation with Pydantic. Authentication: API keys, tokens, OAuth at a working level. Consuming third-party APIs, pagination, retries, rate limits, backoff. Week 4 Linux, SQL and the command line: shell fundamentals, processes, environment variables, file permissions, SSH, cron. SQL: SELECT, joins, GROUP BY, subqueries, indexes. Connecting Python to Postgres. ★ Skills you leave withProduction PythonpytestGit & code reviewREST APIsSQLLinux & shellPostgreSQL
# What you'll ship
Real projects. Real users. Real receipts.
Milestone 2
A deployed, containerised, multi-user application with authentication, background jobs and a working UI. No AI yet — deliberately.
Phase 1 · Weeks 5–8
Milestone 3
A document-processing service with schema-validated output, a test set, and a measured accuracy figure the student can defend.
Phase 2 · Weeks 9–12
Milestone 4
A production-quality RAG system over messy documents, with citations, permission filtering, and a written retrieval quality report.
Phase 3 · Weeks 13–16
Milestone 5
An agent completing a real multi-step business workflow, with permissions, approval gates, an audit trail and a trajectory eval suite.
Phase 4 · Weeks 17–20
Milestone 6 (Capstone)
An earlier project taken fully to production — evals in CI, monitoring, cost analysis, security review, deployment pipeline, and documentation another engineer could take over.
Phase 5 · Weeks 21–24
Direct fits — the same job under different names
AI EngineerAI Software EngineerGenAI EngineerLLM Application EngineerAI Agent EngineerAI Product EngineerConversational AI Engineer
# Pricing
One fee. No surprises.
Standard programme fee
₹3,00,000
The full fee before the current enrolment offer.
50% OFF
Limited offer fee
₹1,49,999
You save ₹1,50,001
Offer valid till 31 August
✓ Includes internship placement
✓ 12-month placement support
✓ All exam vouchers and cloud credits
✓ EMI from ₹6,250/mo for 24 months
# FAQ
Answers to the real questions.
AI Engineering · Current fee ₹1,49,999Pay in EMI 8 seats left · March 3, 2026