Somesh Raju

Background

About Somesh Raju

For me, a machine learning model isn't finished when it hits 95% accuracy in a notebook. It's finished when it's a robust, scalable service running in production, actively solving a problem or creating value. That's the principle that has guided my career, blurring the lines between Data Scientist, MLOps Engineer, and Software Developer. I thrive on end-to-end ownership, taking a project from initial data exploration and prototyping all the way through to building the CI/CD pipelines, containerizing it with Docker, and deploying it as a live API. It’s this full-stack approach that allowed me to lead the creation of an award-winning advertising solution on AWS SageMaker and to engineer a GenAI chatbot that empowered our marketing teams and cut ad-hoc requests by 40%. At the core, I'm a problem-solver who uses data as a tool. Whether it's building a churn model that generates an extra $20k/month for an OTT platform or an AI scraping engine that improves audit efficiency by 75%, my focus is always on the tangible impact. I'm always eager to connect with others who are passionate about building intelligent, data-driven products that work in the real world. Let's talk about tackling complex challenges. Key Areas: • Data Science Modeling: Statistical Analysis, A/B Testing, Forecasting, Experimentation • GenAI & LLMs: LangChain, Bedrock, RAG, Fine-Tuning • MLOps & Production ML: AWS SageMaker, Vertex AI, MLflow, Docker, Terraform • Data & Software Engineering: Python, SQL, Airflow, PySpark, API Design