Parth Verma

Machine Learning Engineer focused on building production-grade Gen AI systems, with hands-on experience building LLM-powered solutions that improve customer experience, automation, and product intelligence at scale leading to increase in Automation Rate and CSAT for Coinbase saving Millions every year. Currently at Coinbase, where I work on LLM-driven CX intelligence and agentic AI systems—from detecting recurring product issues and automating issue routing to enabling chatbots to resolve customer queries autonomously. My work involves model evaluation and migration, prompt tuning, workflow orchestration, and balancing latency, cost, and accuracy in real-world production environments. Previously, I’ve built AI-powered applications across startups, including autonomous agents, ML-integrated platforms, and data-driven systems. I enjoy working at the intersection of machine learning, backend systems, and product impact, turning noisy data into reliable, actionable intelligence. Technical focus areas: Machine Learning & AI: Large Language Models, NLP, Transformers, Agentic AI Engineering: Python, C++, Airflow, LangGraph, PyTorch, Google ADK Data & Cloud: AWS, Vertex AI, Databricks, Snowflake I was a Computer Science student at IIIT Lucknow and have previously led ML-focused learning initiatives within the campus community. Always open to conversations around LLMs, applied ML, and scalable AI systems.