AI Engineering — Real-World Projects
15 portfolio-ready AI projects spanning ML, LLMs, RAG, agents, and production deployment.
Pick at least one from each level. The advanced projects (RAG, agents, multi-modal) are exactly what product companies like Sarvam AI, Krutrim, Zepto, and CRED are hiring for. Deploy every project on Railway with a live demo URL.
Train a Random Forest on a labelled email dataset. Serve predictions via FastAPI. Deploy to Railway. Accuracy >95% target.
Regression model on study-hours, attendance, past scores → predict final grade. Interactive prediction form with Chart.js.
Parse PDF resumes, extract skills with TF-IDF/regex, score against a job description, rank top-5 candidates automatically.
RAG over company FAQ docs. User asks a question, FAISS retrieves relevant chunks, LLM answers with citations. React chat UI.
Fine-tune or embed article text, binary classifier (real/fake). Explain which phrases triggered the flag. FastAPI + browser extension.
Collaborative filtering + content-based hybrid. User rates 5 movies → get 10 recommendations. Cold-start handled with genre preference.
Input a YouTube URL, fetch transcript via API, chunk and summarise with LLM, output bullet-point summary + 5 key timestamps.
Paste Python code, LLM reviews for bugs/style/performance, returns structured JSON: issue type, line number, suggestion, severity.
Scrape or use dataset (house/car/gold). Train regression + feature importance. Interactive sliders → live prediction. Deployed API.
Upload PDF (invoice, contract, report). Extract key fields with LLM, answer questions about the document, export structured JSON.
Describe what data you want in plain English → LLM generates SQL → runs on your database → returns result table. Handles joins, aggregates.
CNN trained on PlantVillage dataset. Upload leaf photo → identify disease + treatment advice from LLM. Mobile-friendly FastAPI.
RAG over RBI guidelines + income tax rules. Chat interface: ask about 80C deductions, home loan tax, NPS benefits. Never hallucinates rules.
Input syllabus PDF + exam date. Agent breaks topics into daily plan, generates quiz questions per topic, tracks progress, adjusts schedule.
Upload product image + raw description. AI extracts structured fields (name, category, specs, price tier), generates SEO title and tags auto.