Full-stack engineer working across backend, AI, and the maths underneath. Based in London.

BSc Honours in Computer Science & Mathematical Statistics (UCT, first-class). Five years shipping production systems across fintech, proptech, and AI platforms — from solo MVPs to live products serving thousands. Strong backend focus, with a thesis on evolutionary AutoML for financial time-series forecasting.

I care about systems that are correct under load, models that are honest about their uncertainty, and code that someone else can read six months later.

Servvio — a software consultancy delivering production backends, APIs, and AI-powered platforms for startups and SMEs. I architect and ship the systems end-to-end: Node.js / Next.js / Supabase, layered design, real test coverage, and AI-assisted workflows (Claude Code) that keep delivery fast without sacrificing architectural consistency.

Recent engagements:

• ChavrutaAnytime (TorahAnytime) — AI-powered social matching platform connecting Jewish learners worldwide with compatible study partners. Owned backend architecture, API design, database schema, and the matching logic; production-ready in three months.
• Surety — production MVP for an Israeli prop-tech startup: secure rental deposit management with automated workflows and full audit trails. Delivered end-to-end under tight time and budget constraints.

• Passionstocks — co-founded and built a web platform for trading simulated "stocks" tied to social media creators. Full backend in Laravel: trading logic, user wallets, payment flows, REST APIs.
• Squaretalk (intern) — shipped a REST API integration for buying and selling virtual phone numbers; adopted by thousands of users in production.
• Freelance, 2015–2019 — 30+ bespoke web applications and internal tools for clients (PHP, JavaScript, Python).

Evolutionary Algorithms for Automated Machine Learning (BSc Honours thesis, UCT, 2021) — built an end-to-end AutoML system in Python (PyTorch, NumPy) using evolutionary algorithms to autonomously design and optimise neural network architectures. Applied to financial time-series forecasting, with results competitive against hand-tuned industry baselines.

Languages — TypeScript, JavaScript, Python, SQL, PHP Backend — Node.js, Next.js (API routes), Laravel, PostgreSQL, Supabase Frontend — React, Next.js, React Native, Tailwind AI / ML — PyTorch, NumPy, Pandas; LLM application work (RAG, embeddings, vector search, agentic workflows) Infra — Docker, GitHub Actions, Vercel

• Email — adamlewison@outlook.com
• LinkedIn — linkedin.com/in/adamlewison

Open to interesting conversations about backend systems, applied ML, and early-stage engineering problems.