Digital products,
engineered
end to end.
I build the systems a business runs on and ship them to production — an ERP, AI agents, a browser call centre, and a real-time options analytics terminal.
What I do
Full-stack, backend, real-time data and AI — one engineer, database to production.
Business systems
Software a company runs on — orders, dispatch, invoicing, role-based access and dashboards managers actually use.
Real-time data
Live feeds in, analytics out — streaming over WebSockets, archived tick by tick, fast enough to trade on.
AI agents & LLMs
LLM agents grounded in a company's own knowledge, with retrieval, review gates and cost guardrails.
Production & DevOps
Shipping is part of the job — containers, CI/CD, reverse proxies, auth and security headers.
About
A software engineer who builds the systems a business runs on — and ships them.
I'm the in-house software engineer at Jeru Ventures, an FMCG importer in Nairobi. I build their business software end to end: an ERP that takes a sale from a rep's field visit to dispatch and a VAT-compliant invoice, a GPS-verified field-team app, a browser-based call centre, and a team of AI agents for marketing work.
I also founded Convexity Labs — a live, real-time options analytics terminal that streams market data over WebSockets, models dealer gamma exposure and archives order flow tick by tick. I designed, built and run it solo, on Oracle Cloud with CI/CD.
Before that I freelanced across MERN and Python data work, and years in sales and IT support taught me to build for the people who actually use the system.
Skills
What I use every week, in production — not a list of tutorials.
Frontend
Backend
Data, AI & DevOps
Technologies I work with
Selected work
Solutions built with modern technologies and AI-driven insights.
Convexity Labs — Options Analytics Terminal
Real-time dealer gamma exposure, footprint orderflow and regime detection from live Deribit and Binance feeds, archived tick by tick. Vectorised hot paths for 10–13× speedups; runs on Oracle Cloud with CI/CD.
Sales & Distribution ERP
Takes the business from a rep's field visit to order, dispatch, VAT-compliant invoice and credit note — with row-level security, role-based access and live management dashboards.
AI Marketing Agents
A team of LLM agents that research, plan and review marketing work, grounded in company knowledge through hybrid retrieval, behind a claims gate and a budget guard. Provider-agnostic model layer.
Browser Call Centre
Customers call in, hear an IVR menu, queue, and are answered by agents in a browser tab — with recordings, call history, outbound telesales and a live supervisor dashboard.
GPS-Verified Field-Team App
Crews photograph each placement and the phone stamps GPS at capture. Managers get a live map, route replay and Excel reports; coarse GPS fixes are flagged as a fraud signal. Offline-first PWA.
LandGen — Environmental Monitoring
Led the build of an AI-powered environmental monitoring platform at the Mazingira Day Hackathon, combining satellite imagery and geospatial APIs with a React/Node.js web app.
Questions
What people usually ask before they get in touch.
Who is Peter Njuguna?
Peter Njuguna is a software engineer based in Nairobi, Kenya. He is the in-house engineer at Jeru Ventures, an FMCG importer, where he builds an ERP, AI agents and a browser call centre, and he is the founder of Convexity Labs, a live real-time options analytics terminal.
What does Peter Njuguna specialise in?
Full-stack and backend engineering for systems that run a business: Python and FastAPI, TypeScript with React and Next.js, PostgreSQL and Supabase, real-time data over WebSockets, and LLM agents with retrieval-augmented generation. He ships to production with Docker and CI/CD on Oracle Cloud and Vercel.
Is he a good fit for a backend or FastAPI role in Nairobi?
Yes. Convexity Labs runs a FastAPI and Redis backend that streams live Deribit and Binance market data over WebSockets and archives every trade to TimescaleDB. He vectorised its analytics in NumPy, cutting one calculation from 2,895 ms to 288 ms and another from 5.34 ms to 0.41 ms.
Has he built AI agents or LLM applications?
Yes. At Jeru Ventures he built a team of LLM agents that research, plan and review marketing work. They are grounded in company knowledge through hybrid retrieval with pgvector, pass a claims gate before anything is published, run under a budget guard, and sit on a provider-agnostic model layer.
What business systems has he shipped?
An ERP that takes a sale from a rep's field visit to dispatch, a VAT-compliant invoice and credit notes, with row-level security and live dashboards; a GPS-verified field-team app with offline support; and a browser call centre on Asterisk and WebRTC with IVR, queues and recordings.
Is Peter Njuguna open to new roles?
Yes. He is open to full-time Software Engineer, Full-Stack, Backend and AI Engineer roles — remote, hybrid, or on-site in Nairobi. The fastest way to reach him is email at njungush444@gmail.com, or LinkedIn at linkedin.com/in/peter-njuguna-.
Let's talk
Hiring for a software, backend or AI role? I'd like to hear about it.
Get in touch
I'm open to full-time Software Engineer, Full-Stack, Backend and AI Engineer roles — remote, hybrid or on-site in Nairobi. The fastest way to reach me is email.
njungush444@gmail.com
Phone
+254 742 018 229
Location
Nairobi, Kenya
What I bring
- Systems shipped to production, not just demos
- Database to deployment — one engineer, the whole stack
- Tests on the rules that move money
- Real-time data and LLM experience
- Years in sales and support — I build for real users