• Led and mentored global data engineering/science teams for the extraction of diagnostic codes and claims from health records with strong focus on data quality and traceability.• Built a problem-solving oriented and customer-centric culture with KPIs, clearly defined goals, ownership, and risk management (test-driven development).• Applied large language models to automated code and documentation generation.• Spearheaded MLOps roadmap implementation.--- Data engineering ---• Reduced models deployment time from 7 to 1 day by implementing dynamic routing, auto-replay, snapshots/data quality control, declarative configuration, latent documentation, data observability, data contracts, and CI/CD automation.• Scaled processing throughput from 2.5 to 46 medical records/sec using scalable architecture: Kafka, Spark, batching vs. streaming, Airflow, NoSQL/MongoDB, lambda architecture, Databases, Kubernetes.• Implemented a Neural Architecture Search library and Tensorization of datasets for data scientists: Spark, Parquet, PyTorch.• Orchestrated 58 successful releases to production, AWS/EMR/RDS/S3 leveraging agile development, pre/postproduction tests/updates: A/B, Thompson sampling.--- Machine learning ---• Evaluated clinical notes embedding techniques: Word2Vec, Autoencoders, BERT.• Developed a 74.9% accurate transformer model for health insurance claims prediction.• Applied HIPAA compliant data augmentation, reducing training set imbalance by 28%.

Mobile game discovery and monetization platform• Designed and deployed ad targeting models, processing 25,000 ad campaign requests/sec.• Implemented Feature hashing, class balancing, and online exploration using Spark/NoSQLCourse: Functional program design in Scala

Dell big data analytics enable business users to rapidly transform structured and unstructured data into analytic insights.• Led unstructured data transformation for Dell big data analytics platform using databases and emerging distributed computing technologies.• Implemented Feature hashing, class balancing, and online exploration.

Cloudweaver's Application Defined Network solution allows developers and operation engineers to monitor, manage and optimize cloud networking resources• Led agile development, testing, deployment of a SaaS platform for resource optimization. • Applied machine learning for bottleneck remediation on AWS. (Company acquired by F5 Network)

24x7 Content solutions helps enterprises and public institutions to search and navigate large amount of documents.• Designed semantic search platform, patented semantic topology extraction design.

Zvents, the leading producer of entertainment guides online, helps more than 10 million monthly users find local events, and helps thousands of event advertisers increase attendance at their venues.• Reduced production issues by 40% and enhanced traffic conversion through data mining.(Company acquired by StumbleUpon as part of eBay)

OnMeta provides interactive advertising solutions for reaching targeted audiences in digital media• Built web analytics solution, focusing on web demographics targeting.

Resonate’s traffic management solutions maximize application delivery and availability.uilt web analytics solution, focusing on web demographics targeting.• Restructured product development, QA, support, and IT operations with enhanced SLA.• Drove customer retention rate to 94% through test-driven, agile development.• Directed offshore engineering groups, reducing costs by 40%.(company acquired by private investors)

Built up the engineering department from 5 to 40 engineers to develop and maintain a hosted B2B procurement SaaS on the market. Responsbile for $1M quarterly budget.(company acquired by VerticalNet)

Led the development of management console, pRISM for debugging and profiling embedded systems (C/C++, pSOS embedded OS, Windows SDK)

Developed process and emission control software for power utilities. Applied linear programming and Neural Network to predict emission of pollutants (EPRI contract).=> acquired by General Electric - Energy division

Created and supported a C++ library for scientific computation: Statistics, Calculus, Numerical Analysis, Simulation and Optimization.