AI Engineering
Ship LLM systems that survive contact with real users.
- Duration
- 8 weeks
- Commitment
- 6 hrs/week
- Level
- intermediate
- Modules
- 4
Most teams can get an LLM demo working in a weekend. Almost none can tell you whether the thing got better or worse after last Tuesday's change.
This cohort is built around that gap.
By the end
What you’ll be able to do
Retrieval, evaluation, and inference economics for engineers who need production-grade AI, not another chatbot demo.
- 01Design a retrieval pipeline you can defend with numbers
- 02Build an eval harness before you build the feature
- 03Cut inference cost without quietly cutting quality
- 04Debug a failing agent from traces rather than vibes
Format
Schedule
Next cohort
Prerequisites
- Comfortable in Python and one typed language
- Have shipped and operated at least one HTTP service
- Basic grasp of vectors and probability
Curriculum
4 modules, in the order they bite
The eval harness comes first
- Task definition
- Golden sets
- LLM-as-judge and its failure modes
- Regression gates in CI
Retrieval that actually retrieves
- Chunking strategies
- Hybrid search
- Rerankers
- Measuring recall honestly
Inference economics
- Token accounting
- Caching layers
- Batching
- Model routing and fallbacks
Observability for probabilistic systems
- Tracing
- Drift detection
- Incident response when output is non-deterministic
- Task definition
- Golden sets
- LLM-as-judge and its failure modes
- Regression gates in CI
- Chunking strategies
- Hybrid search
- Rerankers
- Measuring recall honestly
- Token accounting
- Caching layers
- Batching
- Model routing and fallbacks
- Tracing
- Drift detection
- Incident response when output is non-deterministic
Who teaches it
Pavan Jegurupati
Taught around systems that actually ran, including the ones that broke. Every session is live, and the project you build is reviewed by the person who wrote the module.
Built on this track
What people finished the cohort with
Pricing
One price, one cohort, no upsell
Everything the cohort includes is in the price. There is no premium tier, and nothing is held back for one.
- Every live session, plus the recordings
- A cumulative project reviewed by the instructor
- Written feedback on your project at two checkpoints
- Cohort channel access that stays open afterwards
AI Engineering
₹37,999
- Status
- Enrolling now
- Cohort
- 3
- Runs
- 5 Oct 2026 – 30 Nov 2026
- Sessions
- Mon & Thu, 7:30-9:00 PM IST
- Seats
- 12 of 40 left
Waitlist
Hear about the next cohort first
Cohorts are small and dates move. The waitlist gets the date and the early-bird window before the site does.
Questions
Before you enrol
Anything not answered here is worth an email — the reply comes from the person teaching it.
Ask something elseNo. Prompting is one lever among many, and rarely the one that decides whether a system holds up. We spend most of the time on evaluation, retrieval quality, and cost.
Roughly six hours - two live sessions plus project work. The project is cumulative, so falling behind compounds.