The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
Designed a modular ML architecture with dedicated layers for APIs, feature engineering, inference, persistence, testing, and artifact management to support reproducible training and scalable prediction workflows.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Trending
retrofuturism
Share your favorite designs from the past exploring the distant future. Or your own take on the aesthetic 🚀
halftone
Dots. Texture. Grain. Let's see your designs using the halftone technique.
Midjourney
Creatives are showing off what they're making with Midjourney.
colorgrade
Video editors sound off: How do you enhance the mood of your video in color grading?
Related posts
Just published: Strategy and Copy Agent.
Most AI copy tools optimize for generation. I wanted to solve the harder problem:
How do you make AI-generated strategy evidence-backed, reviewable, versioned, and safely approved?
This case study breaks down the system I built around:
Evidence → Strategy → Copy → Review → Approval → Handoff
Including role controls, immutable revisions, deterministic checks, and exact-version approval.
If you’re building production AI systems where trust and control matter as much as output quality, this one is for you.
Keep going
Open to new work — full-time or contract.
I'm a full-stack engineer with 6+ years of experience, working with TypeScript across the stack. Most of that time has been spent embedded in a small number of products for years at a time, rather than rotating between short gigs — which means I've owned things long enough to live with my own decisions.
What I build:
AI/LLM products — tool-calling agent workflows that act on results autonomously, retrieval-grounded generation over large corpora, structured outputs with validation, and evaluation against real outcomes. Built in-house, with the model provider behind a service boundary so routing stays swappable.
Payments — Stripe Connect and Subscriptions in production: recurring billing, plan changes, usage-based charges, marketplace payouts, and webhook-driven reconciliation so the ledger stays accurate when events arrive late, twice, or out of order.
Integrations — 15+ third-party APIs including Amazon Seller/Vendor Central, Google, and Microsoft Graph behind one normalized interface, cutting integration time by ~40%.
Real-time & scale — WebSocket services maintaining consistent state for thousands of concurrent users, with reconnection and out-of-order delivery handled explicitly.
Stack: TypeScript · React · Next.js · Node/Express · PostgreSQL · Redis · AWS · Docker · CI/CD. Claude Code and MCP tooling are part of my daily workflow.
If you're shipping something early-stage and need someone who can take a feature from schema to production without a handoff, my DMs are open.
good luck in your job search Ali
Project Title: Quantum Prism: Autonomous Hardware Diagnostics AI
Description:
Designed an autonomous multi-agent AI system that diagnoses complex laptop motherboard faults. The architecture uses Claude Code as the terminal orchestrator, managing sub-agents in self-hosted n8n via the Model Context Protocol (MCP).
Instead of generic AI wrappers, this system relies on a Python FastAPI microservice for computer vision (analyzing boardviews/schematics) and a Qdrant Vector DB for RAG-based hardware theory. It enforces a strict, zero-hallucination diagnostic cycle based on exact physical values (Volts/Ohms/Pins) to navigate from a raw symptom to a precise component verdict.
