Unified Event Detection Pipeline
Apache Flink + Kafka pipeline consolidating 3 ingestion sources into a single platform — sustaining 10M+ datapoints/sec, 40% lower detection latency, zero duplicate-incident noise.
Staff Software Engineer at Cisco ThousandEyes — I design petabyte-scale, low-latency distributed systems and AI-augmented observability platforms that serve 10K+ enterprise accounts.
Currently leading the Events platform — Apache Flink + Kafka pipelines, ML-driven correlation, and GenAI summarization on AWS Bedrock.
Apache Flink
AWS Bedrock
Kafka
Java · Python
01 · About
I'm a Staff Software Engineer with over a decade of experience designing the kind of systems that hum quietly in the background while millions of events fly through them — distributed stream processors, low-latency APIs, and the data platforms that power modern observability.
At Cisco ThousandEyes, I lead the Events platform: a unified detection pipeline on Apache Flink + Kafka that consolidates Cloud Agents, Endpoint Agents, and Connected Devices into one engine — 10M+ datapoints/sec with 40% lower detection latency. I was promoted to Staff in early 2025 for sustained cross-org technical leadership and end-to-end delivery of company-priority initiatives.
These days I spend a lot of time on the intersection of distributed systems and AI — productionizing GenAI summarization on AWS Bedrock (Claude Sonnet 4.5), integrating DBSCAN-based ML clustering for incident correlation, and building anomaly detection that's shared across multiple product lines.
Outside of work I read voraciously (tech and otherwise), follow ML research, and unwind with video games.
Based in
Newark, CA · USA
Speaks
English, Hindi
02 · Experience
Cisco ThousandEyes · Events Team · USA
Cisco ThousandEyes · USA
Oracle Americas · USA
Verisk · PowerAdvocate · USA
Cambridge Semantics · Boston, USA
SAP Labs Private Limited · Bengaluru, India
03 · Skills
04 · Featured Work
Apache Flink + Kafka pipeline consolidating 3 ingestion sources into a single platform — sustaining 10M+ datapoints/sec, 40% lower detection latency, zero duplicate-incident noise.
DBSCAN-based ML clustering microservice (Python sidecar) integrated into the Flink topology — suppressing redundant alerts and improving incident quality for thousands of customers.
Productionized on AWS Bedrock with Claude Sonnet 4.5 — converting raw correlated signals into customer-readable narratives and runbooks. Cut on-call cognitive load and triage time.
Shared service for HTTP, latency, RTT, and jitter signals — improving detection accuracy and reducing false positives. Expanding across multiple ThousandEyes product lines.
Core contributor to Oracle's fully-managed Elasticsearch service across 30+ regions globally — automated cluster provisioning, backup, and patching workflows that cut provisioning time by ~50%.
Full-stack MERN + GraphQL app for creating and booking events. Used Facebook's DataLoader to batch and cache API calls per event-loop cycle, reducing DB reads.
05 · Education
2019 — 2021
Khoury College of Computer Sciences · Boston, MA
Master of Science in Computer Science
GPA: 3.9 / 4.0
Programming Design Paradigms · Algorithms · Foundations of AI · Large-Scale Parallel Data Processing · Building Scalable Distributed Systems · DBMS
2011 — 2015
Shri Govindram Seksaria Institute of Technology and Science · Indore, India
Bachelor of Engineering in Computer Engineering
OOP · Data Structures · Databases · Software Engineering · Cloud Computing
06 · Contact
Always open to thoughtful conversations — distributed systems, AI platforms, mentorship, or just a good engineering story.