• Increased completion of critical sales tasks by 35% by shipping a real-time push notification platform (Go + Kafka); designed idempotent consumers, retry/backoff, and rules-based routing so the right teams acted on the right events fast.
• Cut invoice processing cost 10x and cycle time 50% by building an automated extraction & attribution pipeline on in-house LLM services; unified ingestion → parsing → validation → posting → approval, with confidence thresholds, fallbacks, and finance audit logs.
• Built a ledger accrual and vendor payout system with SFTP based Bank integrations (using Go, Kafka, DynamoDB, Redis) to automate disbursements for 10K+ vendors daily, saving nearly 20 hours of man-power weekly and reducing errors in payouts.
• Improved reliability and security across a microservices stack by optimizing high-latency APIs (profiling + hot-path rewrites) and remediating critical vulnerabilities; introduced safer defaults (authn/z, secret handling) and CI checks (linting, SAST/DAST) to harden production.

• Successfully led a 20% increase in user subscriptions by developing a high-performance news streaming platform, ensuring real-time updates with less than 300ms latency through the utilization of Websocket as a Service and Pub/Sub architecture.
• Enhanced Postgres database operations to achieve a remarkable 3ms latency for reads/writes through indexing and partitioning techniques, optimizing overall system responsiveness.
• Created robust user session management using Redis, enabling seamless authentication and authorization processes while managing over 3,000 daily write operations and 15,000 daily read operations on the database.

• Achieved a 70% reduction in time-to-recovery from failed Risk and PnL jobs by developing a Flask-based API utility with thread pooling, local caching, and RQL for JP Morgan’s in-house trading platform.
• Improved system performance and data retrieval by utilizing an in-house distributed object store with indexing, resulting in a 100% increase in meeting End-of-Day SLAs.
• Enhanced overall system efficiency and reliability by implementing load balancing and optimized data processing frameworks to deliver high-impact business solutions.

• Write, develop, evaluate and document new and modified software programs running on the Openfabric Testnet chain.
• Developing a 3D Human Rotation AI App using HumanNeRF based architecture.
• Research, design and implement algorithms in deep learning for computer vision ang cutting-edge research.

MBON - Marine Biodiversity and Observation Network

• Building and maintaining Darwin Core Visualization python package - PyDwcViz, an open-source python package from scratch, adding 4 modules and 10 functions for providing easy one line analysis functions.
• Brought more than 7 features with a focus on OOP approach, creating 11+ Pull Requests and resolved 3 issues.
• Made it super easy for marine researchers to create diversity indices like Shannon's Index, ES50 (hulbert's index) with one-line functions.
• Plot DwC (Darwin Core) Data on a world map with built-in function, saving time and effort for researcher to focus only on the research and not struggling with writing the code for it.

• Rebuilt the complete pyobis python package - python client to OBIS API, with 100% endpoint coverage, ensuring quality by increasing the unit test coverage to 100% for all modules. github.com/iobis/pyobis
• Implemented 4 new features focusing on code quality, and successfully getting the release of a production package within a span of 12 weeks.
• Resolved more than 27 open issues, and successfully created 30 merged pull requests.
• Extrapolated data visualization tools to bring out research-grade analyses on migration patterns, time-series depth analysis, and species population among others from data captured through all 7 OBIS nodes globally