AI Engineer · Senior Full-Stack Developer · TypeScript, JavaScript, Python

24 years U.S. Army · B.S. in Artificial Intelligence in progress

I build backend and AI systems that people can trust and the next engineer can carry. Most of my production work is on the backend: ETL pipelines, event-driven services, APIs, and the data underneath them. On the AI side I work with LLM APIs, RAG, agents and MCP, with a human signing off on anything consequential.

My father wrote COBOL starting in the 1960s. The lesson came through without ever being said: data integrity is the whole game. If you can't trust the source, nothing built on it is trustworthy either. That's still how I build.

I'm a full-time student. B.S. in Artificial Intelligence at American Military University, expected May 2028. I'm also working through AI Engineering from Scratch, building each piece by hand from the math before reaching for a library.

Joined as senior backend engineer and was CTO within six months. Inherited a monolith that couldn't ship, got the team behind a full rebuild, and put v1 in production in twelve months. I designed and built about 90% of the backend: a hybrid NestJS and Python system that normalizes messy multi-format vendor feeds, with idempotent upserts and alias-driven matching that got more accurate with every feed. → gunkustom.com

Real-time snowfall alerting. MERN, resort-level weather ingestion, SMS and email alerts on user-set thresholds. The hard part was reliability at the edges: cron fetch cycles, 24-hour duplicate suppression, phone validation before the SMS provider, batched reads and writes. → powalert.com

A sanitized rebrand of a real client build. The LLM has no write path to appointments; a human stays in the loop by design. Care assistant with guardrails that falls back to an offline engine, so the demo never breaks. React, TypeScript, Firebase transactions with audit flags, Netlify Functions. → willow-bend.netlify.app

Open source, Apache-2.0. Turns a generative AI initiative into a 5-year cost, benefit and risk projection: payback year, ROI, NPV, IRR and peak funding. Runs in the browser with no account and no backend. 96 regression tests against a source workbook, so a refactor can't quietly move a number. → github.com/MenokoOG/proforma

Open source, MIT. Each failure is a runnable broken.py, a repaired fixed.py and tests that prove the repair: runaway loops, silent tool failures, prompt injection, memory poisoning, a rubber-stamp verifier. 87 tests, no API key required, a written lesson per failure and a code-review checklist. → menokoog.github.io/agent-autopsy · github.com/MenokoOG/agent-autopsy

Static time and space complexity estimator for Python, built for agents. Per-function Big-O with confidence, evidence, and the unknowns it can't decide. CLI, agent tool, or MCP server. v0.1.

My site answers questions about my work with real retrieval: keyword search over a markdown corpus, visible tool traces, and a search-only mode when the model is down. → ljefferson-menoko-site.netlify.app

• AI Learning Lounge: full-stack AI classroom with ETL content ingestion and reading-level simplification.
• AutoForge Lab: containerized crawling pipeline (Collector, Extractor, Validator, Store) on FastAPI, PostgreSQL and Docker.

Earlier client work, 2009 to 2015: menokoog.github.io/Past-Web-Projects-for-Clients-main

• B.S. in Artificial Intelligence, American Military University (in progress, expected May 2028)
• A.S., Web Page, Digital/Multimedia and Information Resources Design, American Military University, 2012
• Full Stack JavaScript Software Engineer, Utah Valley University / V School, 2024

• Proficient AI Engineer, full program, Ed Donner (Aug 2026). Six tracks including Agentic and MLOps.
• AWS Generative AI and AI Agents with Amazon Bedrock, AWS / Coursera (Jul 2026)
• Generative AI Software Engineering, Vanderbilt / Coursera (Jul 2026), including Claude Code
• AI Agents with Model Context Protocol, Vanderbilt / Coursera (Jul 2026)
• Google AI Professional Certificate (Jun 2026)
• PCAP, Certified Associate in Python Programming, and PCEP, Python Institute
• Earlier: Google Cybersecurity, ITIL 4, IBM SkillsBuild badges
• Scrimba "Portfolio of the Week," May 2026
• 24 years, U.S. Army: platoon sergeant, combat engineer, master gunner, CI/HUMINT support, Airborne support
• Master rank, ITF Taekwon-Do

AI / ML: LLM APIs · RAG · agents · MCP · prompt and context engineering · evaluation · QLoRA fine-tuning · Ollama

Backend: Python · Node.js · TypeScript · FastAPI · NestJS · Express

Frontend: React · Next.js · Vite · Tailwind

Cloud and DevOps: AWS · Google Cloud · Azure · Docker · Firebase · Netlify · Render

Databases: PostgreSQL · MongoDB · MySQL · Redis

• Reliability over hype
• One source of truth and a clear data flow
• Observable systems, so debugging is reading, not guessing
• Small, modular code the next person can carry
• Reversible steps: every consequential change can be undone
• Humans keep final authority over consequential decisions

Guided by LAHA: Love All Humans Always.

• Portfolio: https://ljefferson-menoko-site.netlify.app
• LinkedIn: https://www.linkedin.com/in/lawrence-jefferson-ii-46497075
• Credly: https://www.credly.com/users/lawrence-jefferson-ii
• Email: menokoog@gmail.com