Rajeev Bhatia Email & Phone Number

San Francisco, California, Us

- Led the development of AI/ML Platform at enterprise scale for Ashley & Ybor customers with AI workloads on LLMs, Computer Vision & Multi-agent, Multi-modal agentic workflows, large scale distributed ML computing infrastructure & deployment at scale.- The technology team at NAX was acquired by Ybor.ai, an entity of Ashley Furniture Industries- Promoted to Sr. Manager in 14 months

Manhattan Beach, California, Us

- Managing a strong team of ML Engineers & Scientists in building enterprise grade AI Solutions.- Building NLP & Generative AI - Science & Engineering products for NAX.- Developing AI/ML Platform capabilities for enterprise customers using a cloud agnostic, multi-modal, multi-agent stack with serverless, scalable and robust architecture and extreme automation. - Leveraging Knowledge Graphs, Document AI and Graph Neural Networks to create NLP Information Retrieval systems.

Seattle, Wa, Us

- Scaled ~300K ML models on production for AWS Data Centers (DCs) alarming using streaming Anomaly Detection algorithms on thermal and environmental telemetry to improve alarm precision and reducing time-to-detection by ~30%.- Authored research papers on Scalable Time-Series Anomaly Detection and Multivariate Anomaly Detection in Amazon ML Conference 2020 and 2022.- Successfully filed a US patent on Anomaly Detection algorithm technique.- Launched Machine Learning (ML) experiments for AWS data center (DC) thermal optimization helping improve realize 2% total energy savings (20% mechanical energy) by leveraging ML optimization models to influence science roadmap.- Deployed massively scalable ML workflows for minute-level time-series data; presented software design docs, artifacts, prototypes and implemented AWS Batch, Step Functions, Lambda functions. - Mentored a scientists and engineers in building ML models and software platform.

Seattle, Wa, Us

- Developed and launched models to AWS Data Centers for thermal mitigation for by leveraging multivariate anomaly detection on near-real-time streaming data.- Conducted research on data center telemetry from electrical and mechanical SCADA systems to identify opportunities to leverage ML models to optimize Temperature Set points, PID Loops, Predictive Maintenance etc.- Wrote several science and executive papers to present research findings to influence near and long-term science roadmap for the team.

Seattle, Wa, Us

- Conducted end-to-end model development, backtesting and evaluation of various ML algorithms such as Isolation Forest, Random Cut Forest, Extended Isolation Forest, Autoencoders, and Median Absolute Deviation based strategies vs baseline threshold-based approaches. - Partnered with SDE teams to launch low-footprint ML models to Apache Flink based architecture to scale thermal ML models built for each rack in the data center telemetry. Successfully deployed production systems to support up to ~100k models per region.

London, Gb

- Developed a data science pipeline for rate indication process using Machine Learning and RPA to automate the insurance premium prediction.- Created Business Development PoCs and PoVs for several large clients in the banking and financial services domain.

Pittsburgh, Pa, Us

Graduate Research Assistant (RA), at PeeX Lab at Heinz College, CMU with Ph.D. students under Prof. Alesssandro Acquisti, working on GDPR research and Analytics Experiments.

Pittsburgh, Pa, Us

- Graduate Teaching Assistant with Professor Alexandra Chouldechova. “Data Mining” relates to the science of discovering patterns involving methods of Machine Learning, Statistics and Database Systems on large and complex datasets.

Halden, No

- Created Deep learning models for real-time object detection for Nvidia TX-1 GPU on-board a Drone (Unmanned Arial Vehicle) for the autonomous navigation and anomaly detection on electric power lines and grids. The model achieved 90% accuracy with a speed of approx. 18 FPS.- Developed and presented a Face Detection model for real-time face detection using a Drone presented at June 1, 2017 worldwide product launch- Presented insights and model recommendations for further development to the team and stakeholders.

Santa Clara, Ca, Us

- As a senior developer in the 1000+ software development team at Globallogic Inc, worked with clients like Zappos and Datacard on Java/J2EE cloud applications. In addition, involved in research projects and developed a smart warehouse management capability.- Received the People’s Choice Award at the “Global POC Challenge”, 2016 organized by Globallogic all global offices for my project ‘Gloria’ – an IoT based Chatbot. The project used NLP (Natural Language Processing), MQTT protocol and Android integration.

New York, Ny, Us

- As a developer in the 300+ people software development department, worked on developing the SaaS cloud application for the Contact Center with features like softphone, payment gateways, and customer relationship management. - Promoted within a year from Software Engineer to Senior level and mentored three team members in development tasks. Awarded quarterly appreciation twice for exceeding expectations and delivering work. - Received one of two annual “CEO Award” for excellent contribution to software team for the year 2013.

Basking Ridge, Nj, Us

- As a developer in the web application team, worked on software development tasks.- Developed UI and server side components using open source technologies.

Redmond, Washington, Us

- Top Performer for year 2010- Awarded Xbox 360 at Tech Ed 2010, Bangalore- Conducted several technology sessions, on Microsoft Technologies covering 10K + students- Speaker at National Seminar on Cloud Computing conducted at Maharaja Surajmal Institiute of Technology, New Delhi

- Worked on 8051 Microcontroller and RF Modules.Learning:Soldering, Embedded C, Assembly Language for 8051, Wireless Communication, Creating hardware and integration with software.

Master'S Degree, Information Systems Management (Ai & Machine Learning Concentration)

B.Tech, Electronics & Communication Engineering

Education record