// remote · india · utc+05:30
Founder, Yuvanya Systems · Distinguished Engineer
I architect backend systems that serve millions of users - and stay fast under load.
15+ years across distributed systems, cloud platforms and ML-powered products. Patented work in content sync, a platform scaled to millions of users, now building next-generation observability tooling.
A hands-on technology practice that takes engagements from architecture through production - backend systems, cloud & DevOps, data engineering and ML integrations, built and shipped, not just advised. Current focus is next-generation observability: self-training anomaly detection, continuous profiling, and telemetry query tooling.
Lead Computer Scientist → Senior Computer Architect → Distinguished Engineer
Drove technical strategy for a job platform serving millions of users. Led the cloud-native migration to GCP (28% lower infrastructure spend), architected the ML-powered recommendation engine over 7M+ listings (+40% relevance), decomposed the platform into 10+ microservices (30% faster deployments, 40% lower p95 latency), and hardened the edge with bot detection and attack mitigation. Mentored a team of 10 engineers.
Owned engineering at a vehicle-services startup and built the backend platform from the ground up - data model, REST APIs and cloud deployment. Set the technical direction and turned product requirements into a running system that took the business from zero to launch.
Engineered the reverse-proxy core's client-server architecture (25% faster responses) and built a virtual browser for secure web isolation - real-time DOM mirroring with CSS/JS/HTML caching.
Invented and patented the audio-fingerprint content-sync algorithm behind Adobe Dhwani (US9392144B2), led development of the Android client, and drove subscription-licensing development for Adobe Licensing Management.
Built a C#/.NET layout engine for high-volume PDF generation (+30% document throughput) and re-engineered the data-parsing layer, cutting file I/O 40%.
Recent platform engineering across observability and data systems, plus career highlights.
Built a self-training anomaly detection engine for alerting: seasonality-aware baseline models precomputed out-of-band, with zero-allocation scoring in the evaluation hot path - no network I/O at query time.
Built end-to-end continuous profiling: high-throughput ingest across 8 runtime wire formats (Go, JFR, Python, Node, Ruby, .NET, PHP, native), dedicated profile storage, a query layer, and flame-graph UI with differential analysis.
Designed and built a unified query language with a transpiler to the engine's native dialects - one ergonomic syntax across metrics, logs and traces, no raw PromQL/LogsQL required.
Architected an ML-powered recommendation and personalized-search platform - Elasticsearch retrieval with data-science ranking over 7M+ listings, feeding a personalization layer for search, recommendations and lifecycle email.
Engineered bot detection and attack mitigation at the Nginx edge - request fingerprinting, behavioral rate intelligence and automated blocking, stopping abusive traffic before it reaches application servers.
Invented and patented an audio-fingerprint content-synchronization algorithm - granted US patent - and led development of the Android client that shipped it.
Deep dives on the systems I build - distributed systems, observability, performance, and the math behind them.
View all writing →Live from the GitHub API - refreshed weekly by a scheduled workflow.
I take on consulting engagements through my company - hands-on architecture and delivery, not slideware.
Content synchronization algorithm - Adobe Dhwani
us patent application US2015/0373231Video synchronization based on an audio cue - Adobe
For Adobe Dhwani · plus Adobe Bravo for MAX 2013
Reverse-proxy performance optimization (-25% response time)
Layout-engine contributions
Published paper (ISSN 1793-8163, 2010) + two national-conference papers on information security
B.Tech with CS minor, 8.95 CGPA - TIT&S Bhiwani