AI/ML engineer building agent memory, retrieval and evaluation systems.
Liverpool, UK · MSc Advanced Data Science & Artificial Intelligence · University of Liverpool
One frontier brain, an elastic workforce.
Leverage sits under the model you already pay for. A repository and a goal become a task graph; an auction hires local, free or subscription models per task; a failed worker leaves a compact checkpoint that a replacement resumes; and nothing counts until the repository's own tests are green. Exposed over MCP to Claude, Codex or Cursor. In one recorded mission it built five modules of a billing SaaS with five hired workers in 42 seconds, whole suite green, paid inference zero.
Live site and playground · Source · Technical report · Built for the RocketRide x SCU Buildathon
Memory that knows what changed, what remains true, and what was never known.
Lacuna is a temporal, provenance-first memory layer for AI agents, built on HydraDB. It keeps every claim tied to the sentence it came from, preserves corrections instead of silently overwriting history, exposes unresolved conflicts, and abstains when the evidence cannot support an answer.
It projects the same evidence contract through a public web workspace, CLI and MCP surface.
Try the live workspace · Source · Evidence · MCP
| Project | What it demonstrates | Links |
|---|---|---|
| Leverage | An intelligence resource manager for MCP hosts: per-task model auctions, cognitive handoff checkpoints, test-verified proofs | Source · Live |
| Lacuna | Temporal memory, exact provenance, contradiction handling and explicit abstention across web, CLI and MCP | Source · Live |
| QueueProof | Cross-source work retrieval with claim-level citations, deterministic prioritisation and approval-gated writes | Source · Live |
| Kodro | Offline robot coding and visual kinematic simulation for pupils, teachers and beginners, with no account required for core paths | Source · Live |
| BhashaFix | A localisation verification harness covering linguistic, technical, visual and accessibility failures across web, CLI, MCP and CI | Source · Live |
- agent memory, context retrieval and evidence graphs;
- evaluation systems that expose failures instead of hiding them;
- MCP servers, CLIs, connectors and cross-surface contracts;
- full-stack AI products with TypeScript, Python and durable data stores;
- reproducible demos, tests and technical documentation.
Languages TypeScript · Python · JavaScript · SQL
Frontend React · Next.js · Vite · Tailwind CSS · Three.js
Backend Node.js · FastAPI · REST · serverless functions
Data PostgreSQL · MongoDB · HydraDB · Redis · vector search
AI systems RAG · agent workflows · MCP · evaluation · provenance
Delivery GitHub Actions · Vercel · AWS · GCP · Docker
I prefer proof before claims: a working route, a reproducible command, a failing test that becomes green, or an evidence artifact somebody else can inspect.
I document limitations alongside capabilities. A system that cannot support an answer should say so clearly rather than generate confidence from missing evidence.
Useful contributions include failure reports, reproducible edge cases, documentation corrections and tightly scoped pull requests. Start with the open issues in Lacuna or inspect its contribution guide.
For AI/ML engineering, applied-AI collaboration or open-source work: