Smart Contract Developer | Python & AI Engineer | DeFi Infrastructure
Software engineer with three years of production experience across smart contract development, Python data pipelines, and AI-powered automation. Previously build institutional vault infrastructure at Orion Finance and trust infrastructure for autonomous agents at Lyneth Labs.
Current Focus: ERC4626/ERC7540 vault architecture, on-chain quantitative analysis, agentic AI systems, and decentralised trust infrastructure for autonomous agent economies.
Co-Founder | Lyneth Labs
May 2025 - Present | Backed by $20K from Venice AI
Building the trust infrastructure layer for autonomous AI agent economies on ERC-8004 registries.
- Designed a Beta-Dirichlet reputation system that models every agent as a probability distribution rather than a static score, with a Global Mixture Model that dynamically clusters agents into behavioural archetypes
- Built a discrete-time dynamical trust engine treating agent interactions as asymmetric virtual games with logarithmic economic weighting and a Zero-Trust Sybil-Filter
- Implemented wallet clustering and anomaly detection using on-chain metrics (balance history, wallet age, gas spend) to classify behavioural archetypes and detect coordinated attack signatures
- Deployed an EigenTrust-inspired graph layer for detecting Sybil cliques and collusion rings with formal security analysis proving manipulation resistance
Smart Contract Developer | Orion Finance
November 2025 - June 2025
Building ERC4626/ERC7540 vault infrastructure for institutional DeFi portfolio management.
- Developed modular vault factory with strategy adapters connecting to Uniswap, Morpho, Aave, Lido, and Yearn with UUPS proxy and upgradeable patterns
- Built Python pipelines using web3.py, pandas, and numpy to trace on-chain capital flows across Morpho Blue markets, identifying loan/collateral pairs, supply caps, and per-market liquidity depth
- Computed Spearman correlation, Sharpe ratios, VaR, and maximum drawdown across multi-protocol vault positions to surface hidden concentration risk
- Wrote Dune Analytics SQL to track vault activity, token flows, and stablecoin settlement patterns (dune.com/orionfinance/sepolia)
- Optimised gas costs across deployment scenarios with comprehensive Foundry and Hardhat fuzz and invariant test suites
Full Stack Developer | Bond.credit
August 2025 - November 2025
Built crypto-native credit infrastructure for autonomous agents.
- Developed Python credit scoring algorithms processing agent transaction histories and wallet behaviour datasets via iExec TEE for privacy-preserving computation
- Built Solidity contracts for credit vaults and agUSD stablecoin minting against agent reputation scores anchored to ERC-8004 registries
Full Stack Developer | 0em Labs
June 2025 - September 2025
Built stablecoin payment gateways and MCP marketplaces using Node.js, Cloudflare Workers, and WalletConnect, handling cross-chain token flows and automated payment pipelines.
Full Stack Developer | Tokunize
March 2025 - June 2025
Built React.js/TypeScript frontend and Django/PostgreSQL backend for fractional real estate tokenisation with ERC3643 security token standards for regulatory compliance.
Final Year Project, 2025-2026
End-to-end Python ML pipeline for automated smart contract security analysis.
- Fine-tuned BERT classifier on 6,700+ Solidity contracts achieving F1=0.996; benchmarked CodeBERT QA span extractor against Qwen2.5-Coder 480B (zero-shot) and 7B LoRA fine-tuning
- Built full Python training pipeline using HuggingFace, PyTorch, and scikit-learn on Iridis HPC cluster (A100/L4 via SLURM) with LOTO cross-validation across 27 vulnerability categories
- Designed three-stage automated audit pipeline: BERT triage → CodeBERT span extraction → LLM root cause explanation via REST API, reducing manual review scope by ~80%
- Discovered two genuine vulnerabilities in Orion Finance production contracts: unguarded
setVault(DoS vector) and missing ChainlinkansweredInRoundcheck — both disclosed and fixed
CodeGene | Encode AI Hackathon 2025 🥇 1st Place Nethermind Bounty | AI-powered smart contract security platform that automatically generates and executes exploits to prove vulnerabilities. Evolved into Lyneth Labs. Live Demo | GitHub
GuardFi | Encode London AIxWeb3 2024 🏆 $3,500 in Bounties | Decentralised insurance using AI risk assessment and Flare cross-chain verification Live Demo | GitHub
AssetXchange | ETH Oxford 2024 🏅 Main Track Winner | No-code platform for real-world asset tokenisation Live Demo | GitHub
Languages: Python, Solidity, TypeScript, SQL, Java, Haskell
AI & ML: HuggingFace, PyTorch, scikit-learn, BERT, LLM API integration, Claude Code, Cursor
Python & Data: pandas, numpy, web3.py, scipy, data pipelines, REST APIs, Dune Analytics SQL
Blockchain & DeFi: Hardhat, Foundry, Ethers.js, OpenZeppelin, ERC20/ERC3643/ERC4626/ERC7540/ERC8004
Backend & Web: Django, Node.js, Next.js, React.js, Cloudflare Workers, PostgreSQL, Docker
Infrastructure: Git, Linux, Azure, SLURM/HPC, TEEs
BSc Computer Science | University of Southampton 2023 - 2026
Key modules: Algorithms, Machine Learning, Distributed Systems, Cyber Security, Databases
Final Year Project: End-to-end Python ML pipeline for smart contract vulnerability detection — fine-tuned BERT achieving F1=0.996 on 6,700+ contracts, with automated LLM API integration for root cause analysis on Iridis HPC infrastructure. Discovered two genuine production vulnerabilities.