Landry Yves Joel S.

I build machine learning and AI systems that run in production, not just in notebooks. I am a data science engineer with hands-on experience in LLM agents and RAG, NLP, computer vision and cloud deployment on AWS and GCP. I work in Python and take a project from the first look at the data to a working, documented API. WHAT I HAVE BUILT • LLM agent with RAG (LangGraph, LangChain, OpenAI, Mistral): it chooses between vector search and web search, scores its own answers, detects hallucinations and retries automatically (15% fewer errors). Deployed on AWS Lambda with FastAPI, DynamoDB and S3. • SMS fraud detection: DistilBERT fine-tuned with LoRA and Adapters, 95% precision, recall and accuracy. Delivered as a FastAPI service with a web interface, in Docker on AWS EC2. • Medical imaging (ultrasound): label-propagation model served on NVIDIA Triton, 0.03 s per image at 95% accuracy, plus image cleaning and inpainting, with a GCP deployment design. • Computer vision for clients: traffic surveillance, chest X-ray classification, image restoration with GANs (SRGAN, ESRGAN). • Serverless NLP API on AWS Lambda for essay scoring (BERT/RNN, Kappa 97%) and an AWS CI/CD pipeline that cut deployment time by 70%. HOW I CAN HELP • LLM and RAG applications: chatbots, document search, AI agents • Custom ML and deep learning models: classification, NLP, computer vision • Fine-tuning of transformers (LoRA, Adapters) • Deployment: FastAPI, Docker, AWS (Lambda, EC2, ECS), CI/CD, MLflow HOW I WORK • I confirm the goal, the data and the success metric before I write code. • You get clean, documented code and a simple way to run it. • Regular updates, and I tell you early if something is not realistic. Send me a short description of your problem and your data. I will tell you honestly if ML is the right tool and what a first version could look like.