Presentation
• Cloud Modernization: Enabled migration of a critical market-risk application from legacy VM infrastructure to a containerized cloud environment despite complex NAS dependencies.
• Cloud Architecture & Modernization: Modernized enterprise platforms across AWS, Docker, Kubernetes, Microservices, Databricks, Kafka and legacy-system refactoring, strengthening scalability and technology adoption.
• Performance Engineering: Transformed a 4-5 hour IPV data extraction bottleneck into a reliable asynchronous and parallel data pipeline.
• Operational Risk Reduction: Automated long-standing LME trade-data processing that previously required a full day of manual business analyst effort every day.
• Platform Optimization: Reduced Sybase ASE processing from 45+ minutes to less than 5 minutes, eliminating process downtime and L2/DBA intervention.
• Automation Impact: Reduced data aggregation time by 60% while enabling sub-second search retrieval across millions of unstructured records.
Background
• Engineered a remote-file database loading approach that bypassed restrictive Network Attached Storage (NAS) dependencies, successfully enabling the cloud migration of a critical data-heavy application from legacy VMs to containerized cloud environments.
• Eliminated a 4–5 hour data extraction bottleneck by replacing synchronous user-interface downloads with a high-throughput, asynchronous Python-based extraction engine leveraging five parallel processing streams.
• Minimized recurring system failures caused by database connection and memory constraints via process-based connection pooling and dataset pre-caching, stabilizing high-volume instrument-level data workflows.
• Spearheaded an event-driven architecture using Apache Kafka messaging pipelines to integrate Java/Spring microservices with Python analytics engines, distributed data stores, and AWS cloud environments.
bottlenecks across trade, position, and cash processing. Awards: YMD (You Make Difference) | Core Value - Openness |
Core Value - Client Focused Delivery | Core Value - Relationship
Eliminated a long-standing LME operational-risk gap by reverse-engineering non-standard London Metal Exchange
trade formats and creating a custom mapping layer outside the platform's rigid ingestion model.
Slashed down Sybase ASE processing time from 45+ minutes to less than 5 minutes by developing a temporary-table
ranking and indexing approach for targeted data subsets.
Resolved trade cross-multiplication issues by engineering a ranking mechanism for similar-quantity trades despite
the absence of native RANKORDER functionality in Sybase ASE.
Strengthened global reconciliation application delivery for 350+ markets across APAC, Europe, and North America.
Mainframe-to-Linux Modernization: Co-led migration of Cash and Trade data repositories from Mainframe DB2 to
DB2 Linux, reducing legacy technology dependency and enabling platform modernization.
calculations and downstream financial-risk processes. Award: Gold Award
Shortened Java-side liquidity calculation time from 45 minutes to 10 minutes by implementing Flyweight-pattern
caching and reusing common instrument data across regions.
Cut database processing from 10 minutes to 2–3 minutes by calculating only missing instruments and introducing a
reusable static repository, creating an enriched data source for downstream financial workflows.
Managed high-volume financial data platforms processing 6M+ trades and 300+ feeds daily across 70+ products,
supporting critical downstream risk, regulatory, treasury, and finance functions.
Rules Engine & Risk Analytics: Architected a trade-liquidity identifier engine with end-to-end auditability, enabling
line-by-line traceability of liquidity rules and Basel 3 haircut calculations.
automation, and high-volume data-processing solutions. Awards: Best Debutant | Earned 2 promotions and a Best
Performance Award | International Exposure: San Mateo, California | Jan 2007 - Mar 2007
Reduced data aggregation time by 60% by engineering high-throughput normalization and consolidation pipelines
using AWK and shell scripting for unstructured eBay market data.
Visual Search Engineering: Selected for an extended on-site engagement with Like.com in San Mateo, CA,
contributing to visual-search algorithm direction for image-based product discovery prior to its acquisition by Google
Enabled sub-second search retrieval across millions of records by building a custom shopping platform using Joomla
and a Sphinx-based search engine.
Led teams of 5-10 resources across multiple concurrent projects while managing application development, delivery,
and client coordination.
improvement initiatives.


