About
Software Developer with extensive experience at an AI/ML startup, specializing in backend engineering, full-stack development, and innovative solutions for fraud detection. Proven track record in reducing costs and improving application efficiency.
Adept at leading projects and enhancing team productivity through effective communication and technical expertise.
When I’m not diving into lines of code or architecting systems, you’ll often find me soaking up the sun with friends, enjoying the great outdoors. I love spending time outside, whether it’s a sunny day at the park or a casual hangout.
And when I’m indoors, I switch gears to competitive gaming. I enjoy the thrill and challenge of games like Overwatch and Rocket League. Despite my best efforts, I’m stuck at Silver rank—must be the ping, right? I’m also unranked in Valorant, but that just means there’s plenty of room for improvement! Gaming keeps my competitive spirit alive and helps me unwind after a long day of coding.
For business inquiries: mhlungusk@gmail.com
Skills
Programming languages: JavaScript/TypeScript (Proficient), Python (Advanced), Golang (Intermediate), Rust.
Frameworks & Libraries: React, Astro.js, Deno Fresh, Express.js, Hono, FastAPI, Zustand, React Flow, Gin.
Tools & Technologies: Go Wails, Neo4j, Fast AI, Supabase, Vercel, Serverless, Deno, MySQL, SQL, MongoDB.
WORK EXPERIENCE
Darkpools AI/ML (02/2023 – Present)
Insurance fraud solutions startup
Software Developer
- Spearheaded the development and maintenance of four major React components, constituting 90% of the desktop application, leading to a 40% increase in presentations and funding proposals.
- Integrated a custom JavaScript dataframe into the Machine Learning suite, completing an 8-month-old stale component in 3 weeks.
- Developed an Optical Character Recognition (OCR) project in 3 days, to extract and analyze supermarket receipts for fraudulent patterns, generating over R70,000 in revenue.
- Led the creation of a media monitoring tool integrating media APIs for security and monitoring during elections in a sensitive region, improving real-time reporting accuracy by 70%.
Lead Backend Engineer
- Enhanced team productivity by 25% by decomposing a monolithic Python server into three services using Deno TypeScript, Python, and Golang.
- Developed a custom TypeScript implementation of the Python Kedro pipeline framework, halting the migration of ML/Deep Learning models from TensorFlow.js to Python, and reducing development time by 50%.
Full-Stack Web Developer
- Reduced load and development time from over 12 minutes to under 1 minute by leading the front-end migration from Webpack to Vite within 5 days of employment.
- Improved application efficiency by migrating from Electron.js to Golang Wails, reducing application size from 700MB to under 80MB and enhancing load and startup speed by 83%.
- Cut server costs by 97% by moving a legacy server from AWS to a Backend as a Service (Supabase), costing $25 per month.