Godson Ajodo

Watches live AI coding sessions, distills corrections into structured constraints, and retrieves the most relevant ones at the start of each new session. Uses a three-layer retrieval pipeline (Rust AST filter under 10ms, semantic scoring, scope-aware reranker) and mines git revert/fix history to seed the constraint library.

A Lightning-native marketplace that routes paid work to AI agents or human workers. Uses a LangGraph orchestrator to plan multi-step jobs, select workers by reputation and cost, hold funds in escrow, and settle payments as work completes.

A local-first context capture tool for the terminal. Records shell commands with stdout/stderr, watches directories for file changes, indexes everything into SQLite, and answers natural language queries with time-decayed ranking. No network calls required.

Implemented Google's MapReduce paper as a distributed Python framework using gRPC for worker-master communication and Docker to simulate a networked cluster. Includes heartbeat-based fault tolerance with automatic task reassignment, a DAG job scheduler for multi-stage pipelines, and a real-time web dashboard.

Built a Rust CLI that encrypts files with AES-256-GCM and embeds them in the frequency domain of images using DCT, making hidden data undetectable to the human eye. Supports text and file payloads with PBKDF2 key derivation and lossless round-trip extraction.

Built an AI-powered autonomous security remediation pipeline with multi-agent architecture (Scout, Scanner, Triage). Automated GitHub workflow: vulnerability scanning, issue creation, patch generation, and pull requests. Applied Redis vector similarity search to match CVEs with historical fixes.

Created a Symptom analysis chatbot to enable non-native English speakers to describe medical symptoms accurately. Leveraged Interswitch ML for text transformation and Hugging Face models, improving symptom-matching accuracy by 40%.

Published open-source toolkit for automata theory: DFAs, NFAs, CFGs, regex to NFA conversion, and Hopcroft DFA minimization. Enabled instant visualization of automata structures.

Designed an interactive climate change simulation using machine learning libraries. Integrated LLAMA 3.2 for AI-powered storytelling. Secured People's Choice Award at NASA Space Apps Challenge 2024.

Engineered a sophisticated chess engine implementing Minimax with alpha-beta pruning, achieving 1500 ELO. Employed Zobrist hashing for board state evaluation.