Available for contracts

AI Pipeline Engineering
& Automation

I build the systems that run themselves. Custom AI pipelines, distributed infrastructure, and automation that scales — without the overhead.

$ python3 pipeline.py --jobs 500 --mode distributed
▶ Cluster online — 7 workers ready
▶ Dispatching jobs...
[████████░░] 250/500 — 0 errors
✓ Complete. 500/500 processed. 0 failures.
$

What I build

End-to-end systems — from model to production.

Custom AI Pipelines

Multi-model orchestration, batch processing at scale. Built to run overnight without babysitting.

Python FastAPI Flux ComfyUI
🤖

Automation Bots

Scrapers, content bots, social automation, scheduled tasks. Resilient, logged, observable — not fragile scripts.

Playwright Python SQLite Redis
🔍

RAG & Semantic Search

Retrieval-augmented generation systems that let users query documents and databases in plain language — grounded answers, no hallucinations.

Weaviate ChromaDB Embeddings FastAPI
🔌

API Development

REST APIs wrapping your models or services. Auth, rate limiting, logging, Cloudflare Workers — production-ready on day one.

FastAPI Workers SQLite Postgres

Real infrastructure

I don't run your jobs on a laptop. I operate a 7-node distributed mesh — M1 Ultra, dual RTX 4070 Ti SUPER GPU towers, and embedded Linux nodes — all connected over Tailscale WireGuard. Pipelines run overnight, unattended, and I check results in the morning.

M1 Ultra Studio128GB unified memory — model hosting & orchestration
Gemini (GPU)RTX 4070 Ti SUPER, 64GB — image generation & training
Persius (GPU)RTX 4070 Ti SUPER, 64GB — parallel processing
4 additional nodesiMac, MacBook, 2× Raspberry Pi
7
Nodes in mesh
128GB
Unified memory
RTX 4070 Ti SUPER
24/7
Always-on pipeline

Recent work

Examples of systems I've built and run.

AI Mesh Governance Platform

Live

Centralised governance API across 7 distributed nodes. Job registry, audit trail for all tool calls, cross-node memory database, Claude-to-Claude messaging, honeypot security monitoring, and a full web dashboard. ~3,000 lines of Python. Runs 24/7 as launchd services.

Stack: FastAPI · SQLite WAL · Tailscale · macOS launchd · vanilla JS

GrokForge Batch Automation

Live

Chrome extension + FastAPI backend that automates Grok image generation at scale. Job queue, pause/resume, concurrent workers, result storage, and a full management UI. Processes hundreds of jobs overnight unattended.

Stack: Chrome Extension · FastAPI · SQLite · Python · Playwright

Distributed Job Orchestrator

Live

Multi-agent orchestration system with parallel agent spawning, state machine with dependency resolution, shared state, per-agent logging, and completion verification. Reusable pattern — any multi-step workflow drops in and executes with full observability.

Stack: Python · asyncio · JSON state · modular agent architecture

Let's work together

I take on pipeline engineering, AI automation, and infrastructure contracts. 100% remote — all work delivered remotely worldwide. Best suited for: building things that need to actually run, at scale, without babysitting.

Response time: usually same day.