B
Bharadwaja O
About
Detail
United States
Résumé
Jobs
verified_user
0% verified
AI/ML EngineerFlatiron HealthDec 2024 - Current (1 year 11 months)• Processed large oncology datasets using Azure ML and Azure Databricks, handling over 12,000 patient records and saving approximately 1,500 manual annotation hours annually.
• Train transformer-based NLP models with Hugging Face and PyTorch to extract clinical entities, reducing around 2,000 annotation errors per quarter and improving patient-level data quality.
• Automate ML workflows using Apache Airflow and GitHub Actions, supporting generative model experiments with Diffusers and cutting retraining and deployment time by nearly 40 hours per week across multiple clinical projects.
• Fine-tune LLMs using Lo RA and deploy containerized inference pipelines with Docker and Kubernetes, keeping clinical Al pipelines running at- AI EngineerTempusMar 2023 - Nov 2024 (1 year 9 months)• Built predictive analytics and time series forecasting solutions on patient datasets using AWS Sage Maker and Apache Airflow, cutting manual processing by 31% and saving over 3,000 operational hours.
• Automated oncology documentation workflows by developing transformer-based summarization models for medical reports and physician notes using BERT, FAISS, and Pinecone, covering 85% of review processes.
• Strengthened reliability of diagnostic models by creating automated retraining pipelines with ML flow and GitHub Actions, reducing SLA breaches by 25%, improving prediction stability by 35%, and lowering maintenance costs by around $95K annually.
• Reduced processing latency by 40% by optimizing model inference and refining - Associate Machine Learning EngineerDeloitteApr 2020 - Jul 2021 (1 year 4 months)• Created predictive maintenance models using scikit-learn and XGBoost for enterprise clients, reducing equipment downtime by 12% and generating around $180K in operational savings annually.
• Constructed BERT-based NLP pipelines for automated invoice and document processing, boosting entity extraction accuracy to 85% and reducing manual review efforts by 37%, accelerating client reporting workflows.
• Optimized ETL workflows using AWS Lambda and AWS Glue, improving data freshness by 30% and reliably processing approximately 200K records daily, enabling timely analytics for cross-functional client teams.
• Monitored high code quality using pytest and Git, reducing bugs by 25%, preventing 120+ critical incidents, and supportin
Education
verified_user
0% verified
- UMaster of Science in Information Technology ManagementUniversity of Wisconsin, Milwaukee, Wisconsin, USAAug 2021 - Dec 2022 (1 year 5 months)
- SBachelor of Technology in Electronics and Communication EngineeringSathyabama University, Chennai, IndiaMay 2017 - May 2020 (3 years 1 month)