Every roadmap for AI careers I see is a wall of forty technologies. Students screenshot it and freeze. The roadmap that got me from an engineering classroom in India to building production AI systems had five stages and one rule: at every stage, ship something you can show. Here is the twelve-month version I give students who ask.
Months 1–2: Python, APIs and one deployed app
Python well enough to read library code. HTTP, JSON and calling an LLM API. Then a small web app with FastAPI and a front end, deployed with a URL. Not an AI app yet — a working one. This is the floor everything else stands on.
Months 3–4: a RAG project with evaluation
Chunking, embeddings, hybrid retrieval, citations and a confidence threshold, over a corpus you care about — the college-documents project is ideal. Write a fifty-question evaluation set and publish the numbers. This single project answers most GenAI interview questions.
Months 5–6: a fine-tuning project
QLoRA on a small model on a free GPU, with a validated dataset and a base-versus-tuned comparison. Learn what changed and what did not. Now you can speak to both halves of the field with evidence.
Months 7–9: ship for a real user
Take one project to ten real users — a club, an NGO, a shop. Add logging, error handling, a fallback, a privacy note. Enter one hackathon with the same discipline. Real users are what separate a portfolio from a repository.
Months 10–12: make it visible and apply
- READMEs that lead with the problem and the numbers.
- Three LinkedIn posts about what you built and what broke; see building in public as a student.
- A one-page resume with projects above education; see a resume for AI roles as a fresher.
- Apply to roles that mention RAG, LLM apps or AI product engineering, including remote ones.
What to ignore
Certificates as a goal, courses without a project, and the urge to learn every framework before shipping one thing. The students who got roles were not the ones who knew the most tools; they were the ones who could show a system working and explain why it was built that way.
GenAI engineer and AI product builder with 2+ years shipping production-grade AI systems: RAG pipelines, fine-tuned LLMs, hybrid retrieval and multi-modal apps across vision, speech and OCR, architected end to end from ingestion to deployment. Leads engineering for SCULT INDIA's 14-member team, founded the 500+ member TechVerse Enclave community and has mentored 200+ students. Three hackathon first prizes: Changethon 2025 (IIT Roorkee), Product Genesis at Vividhotsava 2025 (CSJMU Kanpur) and BYTEBATTLE (MeetKats).
Open to GenAI roles, hackathon judging, mentorship sessions and guest talks at colleges. Email pranjulrathour41@gmail.com.
