Moiz Ali at work.
I'm Moiz Ali. I build AI agents and automations in Tetovo, and I care most about what comes after the demo: evals, guardrails and an off switch.
01 — Bio
I started with React in 2025, building a website for the Karakoram Alpine Club, a mountaineering club in northern Pakistan, and a run of small apps to learn the craft. In early 2026 I moved to automation: an n8n voice agent, then a workspace where creators hand tasks to AI agents.
I work best with small teams who have a repetitive job a person still does by hand: research, outreach, reporting, data entry. I like problems where a wrong answer is expensive, because that is where checks written in code earn their keep.
A project with me starts with the task as it is done today, then a working pipeline on real inputs, measured: what it costs, how long it takes and how often it is right. A person stays in the loop wherever a mistake would matter, and every run is logged so you can see what it did.
02 — Now
- Building Concord, a supervisor that keeps multi-agent runs alive through crashes
- Taking the cold outreach agent to its first real email, and measuring its reply classifier
- Open to projects that take a repetitive job off a small team
03 — Timeline
- 2025Started with ReactA website for a mountaineering club in northern Pakistan, and small apps to learn the craft
- 2025SEEU ConnectA student platform for South East European University, later rebuilt in C# and PHP
- 2026First agentsAn n8n voice agent and an AI-agent workspace for creators
- 2026World Cup 2026 predictorA data warehouse, six features and five models compared
- 2026Agents you can checkThe cold outreach agent and Concord: evals, guardrails and crash recovery
04 — Toolbox
- Models
- Claude · scikit-learn · XGBoost
- Backend
- Python · FastAPI · Docker · TypeScript · Next.js
- Data
- Postgres · Redis · SQLite · Exa
- Ops
- OpenTelemetry · Railway · Supabase · GitHub Actions