Jun 2020 – Jun 2023 · 3 yrs · Hyderabad, Telangana, India

• Delivered features and shared maintenance duties on 'Insights Ingestion' system, constituting multiple Go services which ingest partner-data (100ks of organizations) at a high volume (80M events/day) via Kafka in a fault tolerant manner, into 'Analytics' platform.
• Owned the fan-out work from platform db-upgrades (MySQL5.7 to MySQL8)
• Extended 'Incremental Ingestion' service to consume newer streams via additional topics
• Worked on adding batching to DLQ consumers for faster recovery and lower latency
• Owned feature delivery on 'Eventer' service, which is responsible for user-wise commit statistics for every repository on GitHub (100M+ users, 372M+ repos). This is a Go service, with a MySQL backend, and Kafka for messaging; interacting with a Ruby-on-Rails service.
• Led the design for transitioning into 'Insights Ingestion' and identified bottlenecks.
• Doubled the throughput via batching when transitioning storage layer to CosmosDB.
• Mitigated operational and scaling issues by adding a low-priority queue.
• Owned Azure Test Plans (a component of Azure Dev Ops, having both hosted and on-prem versions) and related products (Test and Feedback Extension, Azure Test Runner), having a C# and MSSQL backend and React/Angular/JQuery frontends.
• Maintained the products with the help of junior engineers and vendors, by addressing customer issues, developing features and bug-fixes, deployment/release tracking and participating in rhythm-of-business activities.
• Trained a new team and transitioned these products to them.

Jul 2019 – Jun 2020 · 11 mos

• Scoped, designed and implemented various backend (and a few frontend) features for 'Assortment Planning', which is a portal geared towards helping planners in the Retail domain. This is a web-app, with an AngularJS front-end and various RESTful backend Java services, using Google Cloud technologies and deployed on the Google Cloud Platform
• Designed and implemented 'Product Hierarchy Service', using PubSub and Datastore. It accepts data of a few million products, processes them to generate ancestor-paths and indexes and serves queries on the products.
• Architected and refactored the 'strategy' feature into a separate 'Strategy Service' using Pubsub, Datastore and Dataflow (for migration)
• Contributed to product development regularly by providing functional, technical and maintainability feedback on upcoming features, designs and code; and identifying technical debt.

Sep 2016 – Jun 2019 · 2 yrs 9 mos

Mar 2013 – Sep 2016 · 3 yrs 6 mos · Bengaluru Area, India

• Designed and implemented failover capability for the ‘Junglee crawler’ using Hibernate, C3PO and OracleRDS in Java, such that when a machine went down, the work allocated to it wasn’t lost and would be picked up by another one.
• Backend has >70 machines, crawls >200 sites a day, visiting overall ~150000 urls a day
• Owned the transition to the failover stack without causing outages, till operational stability was reached, by identifying and fixing bugs
• Proposed the design of a scalable version of a crawler to ‘benchmark’ Junglee, by reusing the ‘Junglee crawler’ (to exploit its multi-threaded capability to crawl more sites in lesser time), while distributing the load of crawling the product urls across different sites
• Scoped, designed and implemented multiple features in Java across multiple business domains:
• Junglee: Backend support within Junglee services for integrating ‘PayWithAmazon’
• Transportation: Multiple features requiring changes across 2 teams and 3 services viz., extensible support for customizable text on labels, rewrite of a pass-through service to support newer interfaces, support for packaging as a new value-added service etc.
• Quickly adapted to different domains and technologies, as evidenced by working on the crawler-backend, then seller-onboarding and then logistics, within a span of 6 months

Jul 2012 – Dec 2012 · 5 mos · Bengaluru Area, India

• Worked on implementing REST-ful web-services and implementing Hadoop jobs, as part of the overall system of ad-campaign planning and lead-generation.
• Worked on SocialFollowup, a service enhancing lead-generation by reaching out to leads via advertiser's social media.

Jan 2012 – Apr 2012 · 3 mos

• Teaching Assistant for EECS 478 (Logic Circuit Synthesis and Optimization)

Sep 2007 – Dec 2011 · 4 yrs 3 mos

• My research interests include network-based anomaly detection, malware analysis and Internet measurement.
• IPv4 Exhaustion: Studies to quantify the effects of the impending IPv4 address exhaustion on the Internet. Primary tools: Python, Unix tools.
• Anomaly Detection: Comparison of detection performance of different approaches to network-based anomaly detection. Primary tools: Python, Unix tools.
• Malware Analysis: Study of evolution of features of malware. Primary tools: CWSandbox, Unix tools.

Aug 2006 – Jul 2007 · 11 mos · Hyderabad Area, India

• Design and development of a tool alongwith a web-service front-end to automate parts of the *nix porting process. Primary Tools: C, Unix tools, PHP, HTML, Javascript.
• Porting of CA products from IPv4 to dual-stack enabled. Primary tools: C, dbx.

Aug 2005 – Jul 2006 · 11 mos · Hyderabad Area, India

• Porting and debugging of core CA components across *nix flavours (Solaris, AIX, HP-UX), and from 32-bit to 64-bit architectures. Primary tools: C, dbx, Unix tools.

Jun 2004 – Jul 2004 · 1 mo · Bengaluru Area, India

• Development of secure communication libraries, for an agent in a grid-computing system, using OpenSSL. Primary tools: C, gdb, Java