M.S. Computer Science (AI) · Senior AI Research Engineer with deep expertise in machine learning systems, cybersecurity architecture, and financial technology. Based in Montgomery, AL.
Selected Projects
Production systems, research publications, and tools built across AI, cybersecurity, and financial engineering.
Student accommodation finder for UNIZIK students in Awka — verified lodge listings, direct landlord contact, and student-friendly pricing built for the Nigerian university market.
Web-based attendance management system for UNIZIK — streamlines student attendance tracking, departmental records, and biometric data authorisation workflows.
Cross-chain DeFi yield aggregator across Base & Solana. Features ERC-4626 auto-compounding vaults, gasless deposits via Pimlico (ERC-4337), deBridge integration, and 37/37 passing smart contract tests.
Production-grade decentralised lending & borrowing protocol (Aave/Compound-inspired) on Ethereum Sepolia & Base Sepolia. Scaled-balance interest model, tiered liquidations, flash loans, and 500+ invariant fuzz test runs.
Gasless transaction relayer for Solana — sponsors user transactions so they never need SOL for fees. Includes a rules engine, JWT auth, key rotation, rate limiting, Prometheus metrics, and reliable rebroadcasting every 2s during congestion.
Enterprise-grade RPC proxy and gateway for blockchain nodes — intelligent caching, WebSocket multiplexing, P2C load balancing, Redis rate limiting, Prometheus metrics, and a Next.js admin dashboard. Kubernetes-ready with Helm chart.
High-performance on-chain event sentinel for Solana — monitors accounts and programs in real-time, decodes Anchor/Borsh data, dispatches batched webhooks (1400+ events/sec), and supports WASM transformer plugins for custom pipeline logic.
CLI tool to scaffold production-ready Solana DeFi dApps in seconds — generates swap, lending, staking, and governance primitives with Anchor programs, TypeScript tests, and a pre-wired Next.js frontend.
Mainnet state snapshot tool for deterministic Solana integration testing — captures live program state (Jupiter, Marginfi, Raydium) and loads it into a local validator. 6–12x faster than devnet, zero flakiness, offline capable.
Production-ready multichain wallet dashboard supporting EVM (Ethereum, Arbitrum) and Solana with ERC-4337 Account Abstraction for gasless transactions, real-time portfolio analytics, and on-chain staking vaults.
Tech Stack
About
I'm Divine Chukwudebelu Ugochukwu — an AI Research Engineer and Computer Scientist with an M.S. from Alabama State University (GPA 3.87), where I was a STEM Excellence Graduate Fellow and published research in IEEE Communications Letters.
At Nexora AI Technologies, I architect production ML systems that sit at the intersection of intelligence and security — fraud detection platforms processing billions in transactions, NLP pipelines parsing millions of compliance documents, and zero-trust cybersecurity infrastructure.
I'm driven by a simple belief: rigorous engineering and applied research aren't opposites. The best systems are grounded in theory, hardened in production, and built to be trusted.
Ugochukwu, D., & Patel, R. (2022). Adaptive DRL for Edge Network Resource Management. IEEE Communications Letters, 26(9), 2201–2205.
Ugochukwu, D. (2023). Machine Learning Approaches to Real-Time Financial Fraud Detection. ICFE 2023.
Contact
Open to senior AI/ML engineering roles, research collaborations, FinTech consulting, and speaking engagements.
+2347016159288