Systems I've shipped — explained the way I'd explain them in an interview.

Off-the-shelf chatbots either hallucinated citations or dropped context across multi-step queries. Faculty needed answers they could trust enough to publish in a course handbook.

Hybrid retrieval (BM25 + dense embeddings), cross-encoder reranking, then a constrained generator that refuses to answer when grounded confidence drops below threshold. State managed through a LangGraph state machine with checkpointing for replay.

• 01Ingestion → chunking with semantic boundaries → embeddings (OpenAI text-embed-3) → pgvector
• 02Hybrid retriever combines lexical and dense recall, top-k = 20
• 03Cross-encoder rerank narrows to top-8 with calibrated scores
• 04Generator runs through LangGraph with deterministic refusal node
• 05Postgres-backed checkpointer for conversation state and audit

Faculty couldn't answer 'how is each team doing?' without manual chasing. Students didn't know what 'on track' looked like. Reviewers received PDFs with no context.

Built a structured workspace around project lifecycle (proposal → milestones → review). Layered AI-assisted progress summaries that read commit history and standup notes to produce honest weekly updates faculty actually read.

• 01Django REST backend with role-based permissions (student / mentor / reviewer)
• 02React frontend with optimistic UI for status updates
• 03Background jobs summarize weekly activity per team
• 04Notification system routes only the changes that matter

Existing aptitude tools optimize for grinding. Students hit plateaus they can't see, so they drill the same topic instead of the next one.

Modeled topics as a directed graph with prerequisites. Per-skill scores update with every attempt; the engine selects the next test set from the frontier of weakest skills with at-risk dependencies.

• 01Topic graph with weighted edges for prerequisite strength
• 02Per-skill score model updated via Bayesian smoothing
• 03Test composer pulls from question bank balanced by skill targets
• 04React frontend with progress visualisations students actually understand

General-purpose social platforms reward outrage and recency. Knowledge communities lose their best contributors to noise.

Topic-first navigation with structured threads, contributor reputation tied to peer endorsement, and zero algorithmic ordering by default.

• 01Topic graph with moderator-curated taxonomy
• 02Thread model with parent / reply / endorsement edges
• 03Reputation system computed nightly from peer signals
• 04Search built on Postgres full-text + topic filters