Choosing a final-year AI project that a recruiter will care about

Walking a room through evaluation criteria
Walking a room through evaluation criteria

Originally published at https://pranjulrathour.scult.in/blog/choosing-final-year-project-ai. That copy is the canonical version and gets updates first.

A final-year project is the one time college gives you months to build something and a panel that has to listen. Most students spend it on a chatbot with no users and a report with no numbers. Here is how I would choose, having reviewed many and having built the projects that got me hired.

Choose by user, not by technology

Start with a person who has a problem you can reach: a department, a local shop, an NGO, a clinic. The college-documents RAG project is a strong default because the users are in the building. A project with ten real users survives any viva question about relevance.

A career session for students
A career session for students

Pick one hard part

Retrieval quality, deterministic extraction, on-device inference, evaluation in a regional language — one genuinely difficult problem you will actually solve. The rest of the system should be plain engineering. Examiners and recruiters both look for the hard part; see the ideas in multimodal LLMs: five student project ideas.

Plan the evaluation in month one

  • A test set with known answers, built before the system works.
  • Metrics that fit: recall and faithfulness for retrieval, field accuracy for extraction, error rates per group for recognition.
  • A baseline to compare against — the naive version, or the tool users have today.

Scope for a demo

One demo path that runs live on the examiner's input. Everything else in the report as future work. A narrow, working system with numbers beats a broad architecture diagram every time, at a viva and at an interview.

Requirements gathering and user flows, on stage
Requirements gathering and user flows, on stage

Document as you go

  • A README that a recruiter can read in three minutes — see a GitHub profile recruiters actually read.
  • Weekly notes on decisions and failures; they become the report's best chapter.
  • A short demo video for people who cannot run it.

Done this way, the final-year project is the first item on your resume and the story you tell in every interview. Done the usual way, it is a PDF nobody opens again. The difference is decided in the first month.

From my carousels
15 Free Tools, No Signup, Every Link
15 Free Tools, No Signup, Every Link, slide 115 Free Tools, No Signup, Every Link, slide 2
15 Free Tools, No Signup, Every Link, slide 315 Free Tools, No Signup, Every Link, slide 4
Pranjul Rathour
Pranjul Rathour
GenAI engineer, Kanpur · 3x first-prize hackathon winner · campus mentor
I ship production RAG pipelines, fine-tune LLMs and build agentic AI products end to end. I lead engineering at SCULT INDIA for a 14-member team and have mentored 200+ students through TechVerse Enclave.
Open to: GenAI roles, hackathon judging, mentorship sessions and guest talks at colleges.
On stage, at hackathons and on campus
Pranjul Rathour, GenAI engineer, Kanpur
Pranjul Rathour, GenAI engineer, Kanpur
Speaking at a MeetKats event
Speaking at a MeetKats event
At VSICS, Kanpur
At VSICS, Kanpur

Comments

Post a Comment

Popular posts from this blog

Forming a hackathon team: roles, skills and the mistake most teams make

Image
At an Integral Startup Foundation hackathon Originally published at https://pranjulrathour.scult.in/blog/hackathon-team-roles-and-formation . That copy is the canonical version and gets updates first. Team TechVerse — Raman Shukla, Saksham Gupta and me — won three first prizes in 2025 not because we were the strongest coders in the room, but because we had stopped overlapping. When I mentor teams before a hackathon, formation is the first thing I fix. In a packed college auditorium The four roles Builder — owns the core feature and the demo path. Usually backend or the AI piece. Interface — owns what the judge sees: the front end, the flow, the seeded data. Storyteller — owns the problem statement, the pitch, the slides and the Q&A plan; also does user research in hour one. Integrator — keeps the pieces working together, manages the repo, runs the demo rehearsals. In a team of three, the builder or interface person takes this. The mistake Five friends who all write R...
Read more

Hello from Kanpur: what I build, and what I'll write about here

Image
On the mic Originally published at https://dev.to/pranjulrathour/hello-from-kanpur-what-i-build-and-what-ill-write-about-here-f2j . That copy is the canonical version and gets updates first. I'm Pranjul Rathour, a GenAI engineer from Kanpur, India. This is the first post on this blog, so here is who is writing, what I have actually shipped, and what you can expect if you follow along. What I build Between February and July 2026, at Next Upgrad Web Solutions, I designed, built and shipped five production AI applications end to end, alone, and deployed all of them with Docker, CI and observability: RAG.NextUpgrad : a retrieval-augmented generation service with hybrid dense plus BM25 retrieval, cross-encoder reranking, streaming cited answers, and a confidence gate that refuses to call the model when the evidence is weak. It runs in about 220 MB of memory, inside a free 512 MB instance. FineTune Studio : fine-tune open models with QLoRA in a browser. Upload a dataset, validate ...
Read more

Reward modeling, explained simply: what it is and why most student projects don't need it

Image
At an Integral Startup Foundation hackathon Originally published at https://pranjulrathour.scult.in/blog/reward-modeling-basics-for-student-projects . That copy is the canonical version and gets updates first. A reward model is trained to score how good a response is, so it can guide further training of another model — the mechanism behind RLHF-style alignment. It's genuinely complex to build well, and most student projects don't need to build one. What it actually requires A dataset of paired responses with human or automated preference judgements — which output is better, and why. A separate training run to fit a model that predicts that preference, before it's ever used to guide anything else. Careful evaluation of the reward model itself, since a flawed reward model teaches the wrong thing to whatever it later guides. When a student project actually needs this Almost never, directly. Supervised fine-tuning or DPO on direct preference pairs gets most projects m...
Read more