Mustapha amine D.

I fix performance and reliability problems in production systems. When something is slow, unstable, or failing in ways nobody can explain, I find the cause with measurements, fix it, and prove the fix holds, from firmware to servers to AI. Most engineers pick one layer and stay there. The problems that actually cost money usually sit between layers, and that's where I do my best work. If any of these sound familiar, we'll probably work well together: - "We keep adding servers and it's still slow." It's rarely capacity. On one SaaS, 74% of request time was sitting in the database. - "Our AI crushed the demo and falls apart in production." I took a document AI from 44% to 95% accuracy in five days without touching the model. - "We're about to lock the BOM." I caught an undocumented MCU routing issue on a neuro-wearable right before manufacturing. - "Something is broken deep in the system and nobody can find it." The bug is almost never where everyone's looking. Some recent work: 1. A control plane for a server fleet of 25,000 machines, with Redfish, IPMI and SNMP behind one API. 2. Offline voice command for a drone system, 13.4 ms from audio to command, no network. 3. Firmware for a neuro-wearable, from pre-manufacturing review to patient-safety interlocks and CI. 4. Predictive maintenance on an 87-unit industrial fleet, with 41% fewer false alarms. 5. An AI agent platform that's handled 1.9M tool calls with under 0.4% hallucination. 6. OTA firmware releases that survive power loss, held at 2% when a build went bad. What the client on that wearable said: "Working with Mostafa has been an absolute masterclass in embedded engineering... His foresight saved us from massive downstream delays and open-ended billing risks." (Full review is on my profile under In Progress work.) How I work: - You talk to me. Not an account manager, not a junior doing the real work. - I usually start with a fixed-price diagnostic, a few days to two weeks. You get the measured cause and a plan, and the fee counts toward the fix if we keep going. - I'll tell you when your problem needs a smaller solution than you're picturing. Not everything needs AI. - I stick around until it works, and after. I also step in as a fractional CTO for hardware and deep-tech startups. Probably a good fit if your business depends on the system and you have a budget from about $1K. Probably not if price is the main filter, or you need a demo rather than something that holds up. Tell me what's broken or what you're building, and what it costs you when it fails. You'll get a straight answer, even if the answer is that I'm not the right person. Things I work with: Performance optimization, root cause analysis, reliability engineering, embedded firmware, BLE, OTA updates, Ambiq Apollo, IoT, edge AI, on-device speech, Redfish, IPMI, SNMP, PostgreSQL, distributed systems, LLMs and RAG in production, AI agents, predictive maintenance, Modbus, fractional CTO.