I build reasoning-centric LLMs from the ground up — data curation, fine-tuning, RL alignment with GRPO, evaluation, and production deployment using vLLM at 10–15× lower inference cost. Before that, I spent 2+ years shipping production LLM systems — fine-tuned them for text generation, conversation, tool calling, entity extraction, and various other use cases, built RAG pipelines with tool calling, and deployed high-throughput inference APIs at scale. I also write technical deep-dives on Medium covering the things most tutorials skip. What I can help you with: 🧠 LLM Internals — fine-tuning, GRPO, DPO, RLVR, curriculum learning, reward design. 🔍 RAG & Retrieval — pipeline design, hard negative mining, evaluation, production readiness. ⚙️ Inference & Deployment — vLLM, LoRA serving, quantization, latency optimization. 📈 GenAI Careers — roadmaps, resume reviews, mock interviews, breaking into AI roles. 💡 Project Reviews — architecture feedback, dataset curation, training strategy. Who I work best with: Engineers building with LLMs who want honest, specific feedback — not generic advice. Students and early-career folks trying to break into AI without wasting months on the wrong things. Anyone who wants to understand why something works, not just that it works. If that sounds like you, book a session — let's get into it.