.png)
Our programmes
Master small language models.
Ship them with confidence.
From prompt design to fine-tuning, from RAG pipelines to edge deployment — structured, hands-on courses that take you from first principles to production.
Available courses
Learn by building, not by watching
Every course combines weekly live sessions, real datasets, and a capstone project you can deploy on day one.
Courses will appear here as soon as they are published.
Can't find what you're looking for? Talk to us about a custom cohort.
How we teach
A four-module path to production readiness
Each module builds on the last. You'll leave with a portfolio of working systems, not just certificates.
Prompt Engineering & Evaluation
Learn structured prompting, chain-of-thought design, and how to measure output quality with gold-standard evals before you touch a single line of fine-tuning code.
- Instruction tuning fundamentals
- System prompt patterns
- Evaluation metrics & datasets
Fine-Tuning Small Models
Take models like Llama, Qwen, and Gemma and adapt them to your domain using LoRA,QLoRA, and full-parameter fine-tuning — on budgets that fit a laptop.
- QLoRA & parameter-efficient tuning
- Dataset curation & cleaning
- Training loops & checkpointing
RAG & Knowledge Systems
Connect your model to proprietary data with retrieval-augmented generation — chunking strategies, vector stores, re-ranking, and hallucination guardrails.
- Embedding models & similarity search
- Chunking & metadata strategies
- Hybrid retrieval & reranking
Deployment at the Edge
Move from notebook to production — quantisation, ONNX export, serving with vLLM and llama.cpp, and monitoring for latency, cost, and drift.
- Quantisation (INT8, INT4, GGUF)
- Serving architectures & APIs
- Observability & cost tracking
What our learners say
Built for people who ship
"I'd been meaning to fine-tune a model for our internal tool for months. The QLoRA module alone cut my trial-and-error time in half. We shipped the prototype within three weeks of the course starting."
Pranav Sharma
ML Engineer, CRED
"The RAG module clarified things I'd been fumbling with for a year — especially the chunking strategy and reranking trade-offs. The instructor pushed us to evaluate everything, not just build it and hope."
Ananya Joshi
Product Lead, Razorpay
"We enrolled our whole infra team in the deployment track. Coming out of it, we replaced our cloud-dependent API with a quantised model running on our own hardware — 60% cost reduction in the first quarter."
Rohan Khanna
Head of Platform, Crio.Do
Frequently asked
Questions we hear often
Not necessarily. Our Prompt Engineering module starts from first principles — how tokenisation works, what a system prompt does, and how to reason about model behaviour. If you've built APIs or worked with any programming language, you're ready. We do assume comfort with Python at a practical level.
All fine-tuning exercises are designed to run on a single consumer GPU (8 GB VRAM and above) or via Google Colab Pro. We provide pre-configured notebooks so you won't spend time setting up environments. If you don't have a GPU, the colab links will get you running in under five minutes.
They are live, cohort-based. Each week has a scheduled session with the instructor, recorded for later viewing. Recordings stay accessible for six months after the course ends. Between sessions, you'll have a dedicated Slack channel for doubt-clearing and peer collaboration.
Yes — a verified certificate of completion is issued once you submit the capstone project and it passes the rubric. We also share a skills transcript that maps what you built to the competencies employers look for, which you can include on LinkedIn or your resume directly.
Absolutely. We offer team and enterprise cohorts with custom schedules, private Slack channels, and a dedicated point of contact. If you're looking to upskill a team of five or more, reach out to us and we'll tailor a programme that fits your roadmap.
Start building
Ready to move past tutorial hell?
Book a free 20-minute counselling call. We'll assess your background, recommend the right course, and answer any questions about time commitment or prerequisites.