Unified Onboarding and KYC

● Worked on Identity, Unified Onboarding, and KYC platforms supporting ~200M users globally.

● Designed and implemented a real-time, event-driven, highly scalable, available, and extensively

monitored policy evaluation system processing Kafka events at scale, executing OPA (Rego) policies,

emitting eligibility decisions, and powering Snowflake and Databricks using an elaborate ETL pipeline

system using Go, gRPC, protobuf, Kafka, Docker, and Kubernetes.

● Led redesign of the Legal Name Change platform by re-architecting APIs and data models, integrating

identity verification, and introducing ML-based (LLM integration) name similarity matching to improve

verification success rates, leveraging Python, Pytorch, LangGraph, PGVector, Ruby, Go, and Datadog.

● Built backend systems for US Tax Onboarding and implemented regulatory-driven offboarding to

remove ~15M inactive users with strong guarantees around correctness, auditability, and safety,

utilizing Go, Ruby, Apache Airflow, and Postgres

● Led critical production initiatives impacting ~150M users on a Code Yellow initiative and

mentored/managed interns while coordinating work across contractors.

Risk Technology and Post-Trade

● Designed and owned large-scale data platforms for risk monitoring, post-trade surveillance, and

compliance, processing petabytes of market and transactional data annually across equities and

crypto markets using Python, multiprocessing, Pandas, Numpy, and shell scripting.

● Built high-quality data models and ETL pipelines powering pricing analytics, index decompositions,

and risk evaluation, leveraging Pydantic, Django, PostgreSQL, and Elastic Search

● Developed data distribution services delivering data to internal systems and external vendors via APIs

and SFTP, achieving up to 20x throughput improvements through parallel processing and system-level

optimizations via websockets, asyncio, and massively parallelized Map Reduce workflows.

● Implemented low-latency and performance-critical components, including a C++ NASDAQ ITCH 5.0

parser and reconciliation safety services deployed alongside FPGA-enabled trading infrastructure.

● Built and maintained internal platforms with a strong focus on reliability, observability, automated

testing, and CI/CD using ELK stack, Java, and Grafana monitoring

● Designed and owned risk monitoring and post-trade surveillance platforms processing petabytes of historical and real-time market data annually, including 100s of terabytes of daily trading volume across equities and crypto markets.

● Built distributed ETL and analytics systems powering pricing models, index decomposition, and portfolio risk evaluation in latency-sensitive trading environments. Achieved 20x throughput improvements and 15x reconciliation speedups through multiprocessing, asyncio-based concurrency, and parallelized MapReduce architectures.

● Developed performance-critical C++ components, including a NASDAQ ITCH 5.0 parser integrated alongside FPGA-enabled infrastructure, reducing feed handling latency by ~35% and enabling microsecond-level deterministic processing. Maintained 99.99% uptime through strong observability, regression validation, and CI/CD automation.

Analyst @ Human Capital Management Division

● Developed and maintained distributed backend services for the firm’s proprietary Human Capital

Management platforms.

● Built notification systems, background verification workflows, required absence, vacation monitoring,

and external vendor data feeds supporting firm-wide operations using Java, Spring, AngularJS, and

Groovy

● Delivered multiple global releases with emphasis on correctness, security, and operational stability.

Technology Division -

I was a part of the Human Capital Management division at Goldman working on designing, developing, and deploying web based full stack applications managing HR related functions like Required Absence and leveraging multiple GS internal tools and technologies.

Tech stack used: Java, Angular 5, Shell Scripting, Groovy