Learn
Pick a track
A track is a short, ordered path to one concrete result. Every step follows the same four parts, and only counts when it leaves something behind — a dataset, a number, a model.
- 1Watch & readA Hindi video, then the paper with a guide.
- 2ConceptsThe few ideas you should be able to explain.
- 3Quick quizTwo minutes to check it really landed.
- 4ActionA small hands-on task with a real result.
-
Track 1 · 9 steps
Fine-tune Qwen for GST
Take a small open Qwen model and make it measurably better at reasoning about Indian GST than it is out of the box — on a free Colab GPU, with the data, the model and the scores published on Hugging Face.
- Measure before you train
- LoRA — what you are actually training
- QLoRA — fitting it on a free GPU
- Format — the chat template
- Synthetic data 1 — generate
- Synthetic data 2 — make it hard, keep it right
- The run — supervised fine-tuning
- Preferences — learn from corrections
- Frontier — reward the reasoning with a rubric recent research
You will use: Unsloth, Google Colab, Hugging Face
-
Track 2 · 9 steps
Faster RAG with Opus and Sonnet
Build a retrieval-augmented GST assistant, measure it honestly, then make it faster and cheaper without making it worse — by sending each question to the smallest Claude model that can handle it and not paying twice for the same tokens.
- How RAG works
- Evaluate it in two halves
- Find where the time goes
- Route — pick the model before answering
- Cascade — answer cheap, verify, escalate
- Compare — a curve, not a number
- Caching, trimming and batching
- Frontier — time-aware retrieval, long context, agentic search recent research
- Put the two tracks together
You will use: Claude API, Opus, Sonnet, your eval set