AI Lab

Where AI Meets
Engineering

A showcase of production systems built through AI pair programming — not AI-assisted, but AI-native from architecture to deployment.

75%+Dev Cycle Reduction
4Production Systems
500k+AI Tokens Processed
274Total Commits
PRODUCTION DEPLOYMENTS

AI-Augmented Systems

View All Projects
88% AI-Contributed

California WARN Pipeline

Live layoff intelligence from raw government filings

AI ToolsAntigravity, Gemini 2.5 Flash/Pro, Cursor, Claude Sonnet 4
ImpactAutomated 100% of data ingestion and alerting
AI Code %88%
Cycle2 days
AI Contribution88% generated or AI-assisted
0%100%
Before AI

Manual Excel download, no monitoring

After AI

Fully automated pipeline, runs twice daily

A fully automated data pipeline for the California WARN Act. Engineered for surgical precision, it transforms raw government filings into live actionable intelligence using ETag caching, MD5 verification, and GitHub Actions CI/CD — running twice daily with zero human intervention.

PythonGitHub ActionsPlotlyETag CachePandas
92% AI-Contributed

Adhan Audio Caster

AI-orchestrated IoT prayer time automation

AI ToolsAntigravity, Gemini 2.5 Flash/Pro, Cursor, Claude Sonnet 4 / Opus 4
ImpactCompressed 3-week build to 4 days · 405k tokens · 8 microservices
AI Code %92%
Cycle4 days
AI Contribution92% generated or AI-assisted
0%100%
Before AI

No automation, manual device control

After AI

Zero-touch IoT orchestration system

An advanced IoT orchestration layer for automated prayer-time notifications. Integrates Raspberry Pi with Sony Android TV via ADB, managing media states and low-level system commands. Built from scratch using AI pair programming—compressing a 3-week engineering cycle into 4 days.

Node.jsADBRaspberry PiShellAndroid TV
75% AI-Contributed

T-Mobile Bill Automation

From manual PDF parsing to fully automated billing pipeline

AI ToolsAntigravity, Gemini 2.5 Flash
ImpactZero-touch monthly billing cycle
AI Code %75%
Cycle~3 days
AI Contribution75% generated or AI-assisted
0%100%
Before AI

Manual Python script, ran per request

After AI

Event-driven E2E billing automation

Originally a manual Python script to parse T-Mobile family plan PDFs and split costs, it was transformed into a fully automated, event-driven E2E system. The Mac Folder Action watches ~/Downloads and instantly processes new bills, calculates splits, and sends Zelle-ready summaries via email.

PythonShellmacOS AutomationZelleSMTP
85% AI-Contributed

Portfolio: bilalahamad.com

Premium AI-native portfolio, built with Next.js + Framer Motion

AI ToolsAntigravity, Cursor
ImpactFull-stack portfolio deployed to production
AI Code %85%
Cycle19 active days
Tokens500k
Commits274
Lines of Code5,600
Saved6 days
AI Agents
AntigravityJul 2025

Google DeepMind

Gemini 2.5 FlashGemini 2.5 Pro
300k tokensFoundation & MVP architecture
CursorApr 2026 – Present

Anthropic

Claude Opus 4.6Claude Sonnet 4
200k tokensModernization, hardening & code review
AI Contribution85% generated or AI-assisted
0%100%
Before AI

Static HTML/CSS resume site

After AI

AI-native Next.js portfolio with analytics

A fully responsive, dark-mode portfolio website built end-to-end with AI pair programming (Antigravity + Gemini). Features glassmorphism design, live GitHub data, Vercel Analytics, Google Analytics with custom events, and an interactive Certifications gallery.

Next.jsTypeScriptTailwindFramer MotionVercel
Dynamic data from ai-metrics.jsonUpdated 2026-04-14

Read the Full Whitepaper

A deep technical analysis on AI-native development methodology, metrics, and the framework I use to measure engineering velocity.