Senior AI Engineer & Automation Architect

Adrien
Bird

Build systems that work exceptionally well,
scale elegantly, and deliver measurable
business value.

8+
Years AI Agency
150M+
Users Impacted
1000x
Process Speed-up
15+
Orgs on METAL+CORE
Innovation Timeline

8+ Years Building
What Others Can't.

Ongoing — 8+ Years Current

AI Agency Founder & Principal Consultant

Lead AI transformation initiatives for enterprise clients. Design and implement automation systems that reduce operational costs. Custom AI solutions that address real business challenges — FastAPI systems, multi-agent pipelines, RAG implementations.

Go Python Claude API OpenAI Redis PostgreSQL
Discord · Spirit Airlines

Senior AI Infrastructure Consultant

Systems supporting 150M+ Discord users: real-time AI moderation, user behavior analysis, content recommendation. Fortune 500 Spirit Airlines: customer service automation, operational efficiency systems, predictive analytics that reduced manual processing time significantly.

Python FastAPI React TypeScript AWS ML Pipelines
Veblen.co · 9 Months

Founding Engineer

Built the account abstraction SDK eliminating crypto complexity from mainstream applications. Transformed complex blockchain interactions into intuitive experiences. Thousands of NFTs minted through the platform with gas-less transactions and social recovery. Brand integration timeline: weeks → days.

Web3.js Solidity TypeScript IPFS ERC-721 Account Abstraction
Various Enterprise Clients

Infrastructure Architect

Financial services fraud detection, e-commerce inventory optimization, healthcare data analytics, manufacturing predictive maintenance. Cross-industry system architecture design and technical leadership for teams of 5–15 engineers.

System Design Cloud Migration CI/CD Security Performance
Services & Products

What I Ship.

Platform Tools

Three Things You Can
Do Right Now.

These tools run live on this domain — no signup, no friction.

Flagship Work

Systems That Moved
The Needle.

01
🤖

NightShift Orchestrator v2

AI Infrastructure Transformation

8 specialized AI agents in a systematic 3-wave strategy. What typically costs $75K–$100K over 12–16 weeks, delivered at a fraction. Evolved into the METAL+CORE Framework, now adopted by 15+ organizations as their AI coordination standard.

15+ orgs adopted
$75K value delivered
02

Safeway Moving Backend

1000x Process Improvement

A 24–48 hour manual load retrieval process was the operational bottleneck crippling revenue capacity. Built a FastAPI system with instant retrieval. From 48 hours to seconds — a 1000x improvement that translated directly to growth.

1000x faster
48h→s retrieval time
03
🔗

Veblen Web3 SDK

Blockchain UX Innovation

Account abstraction SDK eliminating crypto complexity from mainstream apps. Gas-less transactions with social recovery. Thousands of NFTs minted through the platform. Brand integration cut from weeks to days — making blockchain mainstream-accessible.

wk→d integration time
1000s NFTs minted
04
🧠

METAL+CORE Framework

AI Infrastructure Orchestration

Systematic framework addressing 20 critical gaps in multi-agent AI coordination. METAL layer: Memory, Execution, Trust, Audit, Learning. CORE layer: Collaborative, Observable, Reliable, Efficient. Used by organizations across industries.

20 gaps addressed
15+ organizations
Live Build Feed

What's Being
Built Right Now.

FlowEngine workers run 24/7, autonomously completing engineering tasks. This is the actual last task completed.

Connecting to workers...

FlowEngine workers are spinning up. Last task data will appear here.

Open Source & Systems

Infrastructure I've
Built & Shipped.

These systems run in production. Use TaskFlow to build yours.

🎬

GPU Marketplace

GoGinRedisMinIOStripe Connect

A marketplace for renting GPU compute to run AI talking-head video generation (SadTalker, MuseTalk, Wav2Lip). Sellers list GPU nodes, buyers submit jobs, Stripe Connect handles split payments — all orchestrated in Go.

Key finding: Go + Redis pub/sub handles 200+ concurrent job streams with sub-10ms dispatch. Python ML runners communicate via stdin/stdout pipes rather than HTTP — eliminates per-request overhead.
⚙️

FlowEngine Worker

GoClaude Code CLIOllamaFile-lock

Autonomous AI task executor that polls a shared task queue, claims work via atomic file-locking, and executes via Claude Code CLI or local Ollama models. Three workers run concurrently without double-execution — zero coordination overhead.

Key finding: File-based task locking (O_EXCL create) is more reliable than Redis SETNX for local multi-worker queues — survives Redis restarts and requires zero network round-trips to claim a task.
🧠

TaskFlow

GoPostgreSQLOllamaMCP

AI-powered task management SaaS with trained Ollama models for prioritization, time estimation, agent routing, and autonomous orchestration. Ships as two systems: a local MCP for Claude sessions and a remote web server with Cloudflare tunnel.

Key finding: Fine-tuned 0.5B–3B models outperform generic LLMs on structured task operations at 95% lower token cost — task routing, prioritization, and estimation all work offline on-device.
Try TaskFlow →
🖥️

htmx-r

GoHTMXSSEReactive

React-like reactivity without JavaScript frameworks. Declarative state attributes (hx-state-toggle, data-when) drive CSS-powered UI updates. Server pushes state via SSE in htmx-r wire format — zero client-side state management code.

Key finding: Attribute-selector CSS ([data-state-menu="open"] .nav) is faster than class toggling and survives React re-renders. All visual state is in HTML — zero JS class manipulation.
Tech Stack

Built With
The Right Tools.

Languages

Go TypeScript Python Rust SQL Solidity Swift

Frontend

React 19 Next.js 15 HTMX htmx-r TailwindCSS 4 GSAP

Backend & Infra

PostgreSQL Redis Docker Cloud Run Cloudflare Workers Kubernetes AWS

AI & ML

Claude API OpenAI GPT-4 RAG Systems Multi-Agent ChromaDB Ollama MLX
Contact

Let's Build
Something Real.

Have a project that needs serious engineering? Let's talk about what we can automate, optimize, or transform.

720-440-1167 · adrien@adrienbird.com · Denver, CO