Summary
Ayan Dogra is a Senior Data Scientist with six years of experience designing and shipping production-grade agentic AI systems and scalable ML workflows across Google Cloud, Azure, and hybrid environments. He specializes in multi-agent architectures, dynamic tool orchestration, and memory/guardrail strategies that improve accuracy, latency, and cost—having cut agent latency from 8–10s to 1–2s and reduced unnecessary tool execution by 20% in recent projects. His work spans end-to-end MLOps, RAG conversational systems, fine-tuning pipelines, and evaluation frameworks that blended statistical testing with human-in-the-loop feedback to lift model accuracy substantially. Comfortable both in research-adjacent settings and enterprise delivery, he has built systems for market campaign analysis, invoice validation, and complex financial investigations using BigQuery, Vertex AI, Snowflake Cortex, LangChain/LangGraph, and CrewAI. A UW-trained CS practitioner and robotics enthusiast, he brings a pragmatic engineering lens to agent design—favoring orchestration patterns that trade inter-agent handoffs for faster, more reliable single-agent reasoning when it benefits production performance.
6 years of coding experience
2 years of employment as a software developer
Bachelor's degree, Computer Science, CGPA: 9.0/10, Bachelor's degree, Computer Science, CGPA: 9.0/10 at Maharaja Surajmal Institute Of Technology
Master's degree, Electrical and Computer Engineering, CGPA 9.3/10, Master's degree, Electrical and Computer Engineering, CGPA 9.3/10 at University of Waterloo