Agents, retrieval systems, platforms and apps — with the problem, the approach and the outcome for each.
Agentic ML lineage agent
Traces ML and data lineage so teams see what a schema change breaks before it reaches a pipeline.
An agentic tool-calling loop reads DataHub's MCP server, walks the dependency graph, and reports the downstream blast radius of a proposed change in plain language.
Retrieval assistant
Scrapes live web content, indexes it, and answers strictly from that context — flagging contradictions rather than hiding them.
A full RAG pipeline: dynamic scraping, local Hugging Face embeddings into ChromaDB, and Gemini answering under custom conflict-detection logic that surfaces disagreement across sources.
Developer platform
A MERN platform that reviews a file in seconds — bug fixes, optimisation notes and quality feedback.
Built to stay low-latency across multiple languages, with Gemini 2.0 Flash generating context-aware review comments and a React front end for fast iteration.
Cross-platform mobile app
A fitness app generating training and nutrition plans that adapt to goals, level and progress.
React Native on the front, a Django Ninja API behind it, with AI-powered exercise form analysis giving real-time technique corrections and analytics surfacing performance trends.
Tell us what you're building. We'll come back with a scope, a timeline and an honest view of what it takes.