Full-stack & AI engineer with 4+ years shipping production systems. Built a B2B SaaS platform from scratch (3,000+ daily transactions), actively led the frontend modernization of a recruitment platform (6,000+ users), and am building Kobowise, an education-first investing platform that helps young Nigerians learn and practise investing without risking real money.
React, Next.js, TypeScript, Node.js, Python, PostgreSQL, Ruby on Rails. AI/LLM systems. Open to new opportunities.
B2B SaaS platform built from scratch. I designed the initial architecture and built the customer dashboard, product catalogue, ordering, account management, and payment flows, including Stripe checkout with webhook-based order processing, plus Paystack and crypto payments. As usage grew to 3,000+ daily transactions, I worked with another backend engineer on the move to Ruby on Rails and PostgreSQL, and on self-managed VPS hosting with Docker and Kamal.
Initiated and led the frontend migration of an AI-powered recruitment platform serving 6,000+ users from legacy Vue/Nuxt to React, Next.js, and TypeScript, owning most of the implementation. Introduced SSR and validated improved page-load performance through before-and-after testing. Built reusable UI components and recruiter workflows spanning outreach, messaging, sequencing, campaign management, and reporting. Integrated MongoDB-backed data through Apollo GraphQL queries, mutations, and caching, and improved deployment reliability with GitHub Actions.
An education-first investing platform designed to help young Nigerians learn, practise, and understand investing without risking real money. Features guided lessons, NGX company exploration, a ₦500,000 virtual investment wallet, and AI-assisted learning tools. Currently in development.
Turns long forms into quick chat conversations. You answer one question at a time while real inputs appear inside the chat. Includes three ready templates for onboarding, bug reports and support tickets, all powered by one conversation engine with live streaming, strict answer validation, inline editing and progress that survives a refresh. Built with Next.js and TypeScript, animated with Framer Motion and covered by Playwright tests.
Three AI agents passing work through message queues like a factory line. A finder collects headlines, a translator converts them and compliance screens every item, all coordinated over Redis Streams with Docker Compose. Every hop writes to an inspectable trail with timings, failed steps retry then degrade gracefully, and a live dashboard shows runs filling up in real time. Python and Node.js workers behind one REST gateway, zero API keys and zero cost.
Python ·Node.js ·Redis Streams ·Docker ·REST API ·
Built comprehensive sports platform connecting players with local games. Implemented game discovery, instant joining, skill-level filtering, and reservation system across multiple sports using React 19, TypeScript, and Supabase.
Real-time autonomous agent built and deployed in 48 hours for WeMakeDevs Vision Hackathon 2025 — continuous multi-step reasoning, proactive hazard detection, and voice output without human prompting. Agent behavior is governed by a structured instructions file that separates policy from code, enabling rapid iteration without redeployment.
Python ·Gemini Realtime API ·Vision Agents SDK ·Docker ·Railway ·
An in-development AI application that converts natural language questions into SQL queries using local language models. Built with FastAPI, LangChain, Ollama, and Streamlit, with schema-aware query generation and validation to support more accessible database exploration.
“Mark was someone we could count on to get things moving. At PURPLE, he'd pick up whatever needed attention, whether that was a new feature or a last-minute fix. He worked independently, but still thought about how each change fit into the wider product.”
Asaf Remler
Co-Founder & CEO at PURPLE Apps
1 / 9
Let's Build Something
Open to new opportunities and interesting freelance work. My inbox is always open.
