AI agents & automation
Agentic systems that pursue a goal, use your tools and data, and act in your systems, with the evaluation and guardrails to run in production.
AI engineering company
Our model is forward-deployed AI engineering. Nazmi works inside your team and owns the technical path, from the first architecture to the running system.
What Nazmi does
Deep experience in AI agents, automation, and custom model training, backed by production systems, technical writing, and research.
Agentic systems that pursue a goal, use your tools and data, and act in your systems, with the evaluation and guardrails to run in production.
We train, fine-tune, distill, and align open models on your data, then deploy them wherever you want, with you owning the result.
Experience behind Nazmi
10+ years building production AI and machine learning systems. His background includes roles as Principal AI Scientist at In Parallel, co-founder and Chief AI Officer at Resoniks, and AI engineer at Silo AI (now part of AMD), alongside PhD research in machine learning and robotics at Aalto University.
Selected writing
Recent technical notes on AI engineering, models, and production systems.
A Qwen3.5-4B Tau2 retail study where lower loss and train-only prompt gains failed to predict end-to-end agent performance.
Read →Post-training Qwen3.5-9B from 78 to 115 solved tasks out of 220 on a frozen SQL-agent benchmark, matching GPT-5.5 medium for about 50 euros.
Read →A demo proves an agent can sometimes do the job. Production requires evidence it can operate safely, predictably, and recoverably. Six gates every team should pass before letting an agent near real data.
Read →Work with Nazmi
Tell us what you are building, what exists today, and where your team needs help.
Start a conversation or book a 20-minute call →