At Namaah Tech I lead architecture on an AI content-generation platform and write a large share of it: the generation canvas, multi-tenancy with a per-team spend ledger, fal.ai inference integration, and a round of security fixes across the API. I also coordinate a 14-member cross-functional team.
I care about claims that survive checking. When I say a fine-tune cut prompt cost by 91%, or that a model scales recipes with 100% accuracy, there is a rule-based eval behind it. No LLM judges, and the baselines are shown at their best.