Raghavendra
Singh Chauhan
Software Engineer · Distributed Systems · Security Infrastructure
I'm drawn to the deep end — systems that are hard to reason about, problems most engineers avoid, ideas that don't fit neatly into frameworks.
I spent 2.5 years at Deutsche Bank building high-throughput data pipelines and tuning JVMs for heavily-loaded clusters. Seeking intellectual rigor, I took a deliberate sabbatical to attempt India's civil services exam with Mathematics, placing in the top 2% of 800,000+ candidates.
Today, I'm architecting compliance automation at a cybersecurity startup. I build systems that help growing startups verify SOC2, ISO 27001, HIPAA, and GDPR alignment automatically. Mathematics and systems programming both reward rigorous thinking and punish imprecision — moving between these disciplines has made me a more careful, deliberate engineer.
SDE-2
- Architected the compliance automation module on Go / Gin with per-tenant PostgreSQL isolation, eliminating third-party integration dependencies
- Designed an append-only evidence model and pluggable BatchCollector framework across 9+ native sources — CSPM, WAF, DLP, ZTNA, VAPT, Endpoint, SAST/SCA, Logs, URL Filtering
- Built a framework-agnostic control mapping layer — one automated run satisfies controls across SOC2, ISO 27001, HIPAA, and GDPR without duplicating logic
- Designed interface-based abstractions for EventBus and DocumentStore, decoupling core logic from infrastructure and keeping the system cloud-agnostic and swap-ready
Deliberate Sabbatical
Chose to compete for India's most selective examination with Mathematics as optional subject. Ranked in the top 2% of 800,000+ candidates, reaching Stage 2. Deep study of Linear Algebra, Analysis, and Numerical Methods. Rigor, it turns out, is a transferable skill.
Software Engineer — Analyst → Associate
- Built Spring Boot data-ingestion microservices handling 50M+ records/day via MQ and Spark, with 80%+ test coverage and mTLS encryption
- Co-led Gravity modernization — rebuilt data ingestion as an event-driven pipeline, delivering $500K in annual savings
- Owned JVM tuning for a 24-node Hazelcast cluster: analyzed G1GC logs, adjusted heap configuration per server capacity, led post-deployment monitoring during high-load cVA runs
- Led Hazelcast 4.x & Apache Camel 3.x upgrade across a tightly-coupled monolith, averting an 8-week compliance delay
- Isolated data corruption across 160M+ records, resolving a live production outage and averting regulatory impact
- Fast-tracked to Associate (SDE-II) for autonomous ownership of critical microservices
Student Developer
- Built a cell-tracking module using the Viterbi Algorithm with ImageJ, reducing complexity from O(n³) to O(n²)
- Delivered the first production-ready iteration of the tracker
- Distributed indexing pipeline using Virtual Threads and Kafka fan-out for parallel LLM embedding generation
- Vector similarity search via pgvector HNSW indexes and Redis caching for low-latency RAG queries
- Event-driven training platform with async job orchestration and automatic retry logic
- DDPM-based diffusion model for tabular data; Polars streaming for larger-than-RAM datasets
- Merged 2 PRs fixing data integrity issues in US rates export and insider trading display
- Cell-tracking via Viterbi Algorithm; reduced time complexity from O(n³) to O(n²)
IIEST Shibpur
B.Tech, Electronics & Telecommunications
West Bengal, India · 2017–2021