Top-rated AI engineer and full-stack developer with 8+ years building production-grade ML systems, LLM integrations, and voice/agent solutions.
What I do
Build LLM-driven applications:...Read moreTop-rated AI engineer and full-stack developer with 8+ years building production-grade ML systems, LLM integrations, and voice/agent solutions.
What I do
Build LLM-driven applications: RAG pipelines, conversational agents, multi-turn prompt engineering, and retrieval systems (LangChain, LlamaIndex).
Develop AI Agents & voice assistants: CrewAI, AutoGen, Amazon Polly, Deepgram, Rasa, and custom voice/agent pipelines.
Fine-tune & optimize LLMs: PEFT, LoRA, QLoRA, RLHF, DPO using Axolotl, HuggingFace tools, and AutoTrain.
Scale inference & deploy: vLLM, TGI, TensorRT-LLM, AWQ/GPTQ quantization, GGUF, and production deployment on AWS (Bedrock, SageMaker), GCP, RunPod, Vercel.
Data & infra: vector DBs (Chroma, FAISS, Pinecone, Qdrant), dataset generation, evaluation frameworks, Docker, Kubernetes, Redis, and serverless APIs.
Tech highlights
LLMs & frameworks: GPT-4/Claude, LLaMA 3, Mistral 7B, Mixtral, Falcon, Gemma.
Prompting & RAG: multi-turn, few-shot, zero-shot, prompt engineering, LangChain/LlamaIndex.
Inference & quantization: AWQ, GPTQ, GGUF, vLLM, TGI, TensorRT-LLM.
Backend & APIs: Python (FastAPI, Flask, Django), Node.js, Nest.js, REST/GraphQL, AWS Lambda.
Frontend: TypeScript, Next.js, React, Vue, Tailwind CSS, React Native.
Tools: HuggingFace, Transformers, Sagemaker, LangFlow, Flowise, Celery, Docker, Git.
Why hire me
8+ years building end-to-end AI systems and production apps.
Deep practical experience in LLM fine-tuning, RAG, agents, and voice integrations.
Hands-on with inference optimization and cost-effective deployment strategies.
Clear communication, fast delivery, and focus on measurable results.
Availability & credentials
AWS certified, Python certified.
50 hrs/week available for long-term or milestone-based projects.
Ready to design and ship your LLM/agent or voice assistant—let’s discuss scope, latency targets, and deployment constraints.
