Praveen Kumar Gurumurthy

About Praveen Kumar Gurumurthy

ML expert with a decade of experience delivering production-scale AI systems at Microsoft and beyond. Combines deep expertise in ML, GenAI, LLMs, NLP and Social Networks with leadership, product-driven mindset, creative problem-solving, and mentoring to build impactful products, drive organizational efficiency, and influence strategic decision-making across cross-functional, fast-paced environments. Passionate and committed to advancing AI through research, innovation and the responsible development of cutting-edge technologies.

Experience

1. Senior Applied Scientist

   Microsoft

   Mar 2016 — Present · Bellevue, WA, US

   Senior Applied ML scientist and team leader delivering production-scale AI/ML and GenAI solutions at Microsoft. Spearheaded high-impact product features across Shopping, Bing AutoSuggest, and Bing Predicts—driving user engagement, platform intelligence, and cross-team innovation.Microsoft Shopping– Launched Frequently Bought Together, a GenAI recommendation system using few-shot prompting and chain-of-thought reasoning, increasing product page engagement by 11%.– Bootstrapped Shopping Recommendations on MSN Articles, driving 1.4% DAU growth using LLM-based content analysis and LLMs-as-a-judge to estimate defect rates, reducing manual review costs.– Designed personalized Related Searches using user activity signals and LLMs, improving engagement by 6.3%.– Led KPI development (DAU, Personalization Coverage, Product Index Quality) and built real-time monitoring infrastructure.– Boosted CTR by 9% with decision tree-based PClick models; improved recommendation diversity by 36% using MMR reranking.– Co-led graph-based relevance modeling via Shopping Click Graphs; mentored junior engineers and interns.– Served as Capacity Champion, reducing Cosmos storage usage by 50% through optimization initiatives.Bing AutoSuggest– Reduced defective suggestions by 85% by building robust classifiers using deep and engineered features.– Delivered personalized autosuggestions using query embeddings, user profiling, and large-scale clustering.– Deployed production pipelines supporting 1B+ queries per day.Bing Predicts– Built predictive models for elections, sports, and live TV events by integrating entity graphs, web signals, and behavioral data.– Developed high-precision models and features that powered public-facing and internal predictive capabilities.

Education

• Purdue University

  PhD Candidate, Computer Science

  2011 — 2015

• Purdue University

  Joint Master's degree, Computer Science and Statistics

  2011 — 2014

• National Institute of Technology Durgapur

  Bachelor of Technology, Computer Science and Engineering

  2006 — 2010

Skills

• Javascript
• Unix
• Artificial Intelligence
• Computer Science
• Perl
• Data Science
• Sql
• Social Network Analysis
• Linux
• Jsp
• Distributed Systems
• Research
• Core Java
• D3.js
• Java
• Unix Shell Scripting
• Machine Learning
• Web Development
• C++
• Html
• Data Mining
• Shell Scripting
• Software Engineering
• Data Structures
• Information Retrieval
• Matlab
• Operating Systems
• Formal Verification
• Hadoop
• Statistics
• Php
• Python
• Latex
• Algorithms
• R
• Eclipse
• Data Analysis
• C
• Programming