Nov 2020 – Jul 2023 · 2 yrs 8 mos

• Contributed to the complex Managed Volumes product, focusing on data backup and security enhancements. Collaborated with high‑profile clients, including Ikea and Pepsi. Later, delved into the native IBM DB2 database product and mentored interns.
• Designed and deployed a parallelized backup solution. Achieved production readiness in under a month with this user‑centric full‑stack
• feature, collaborating closely with UX, product, and testing teams.
• Spearheaded a team that worked on the UI for two major releases. Introduced DB2 HADR support, and elevated recovery efficiency.
• Collaborated with over 20 users to boost reliability, incorporated preventative tests, and enriched user‑facing documentation.
• Optimized metadata queries, making website loads 20% faster.
• Integrated a Dedupe algorithm, conserving 30% physical data storage.
• Following substantial code additions by the team, product test coverage briefly fell from 70% to 63%. Single‑handedly raised it to 90%.
• Enabled VLAN setup and improved user configuration. Later, collaborated with a parallel team working on similar initiatives.
• In a team setting, designed and implemented Recoverable ranges to boost recovery.
• Formed a team, ideated, and created an acclaimed support tool at a hackathon.
• Led product’s Solaris certification to expand market reach, introduced early backup exit on health‑check failures, and fortified security.
• Implemented essential end‑to‑end tests, bolstering product reliability.
• Commended by leadership for technical expertise, dedication, positivity, and leadership.

Jul 2017 – Nov 2020 · 3 yrs 4 mos · Bengaluru Area, India

• Led Anti‑spam analysis efforts at the Contextual Adverts firm. Managed data ingestion user experience and mentored interns.
• Cut page load delay to 80% and requests to 70% ‑ used Broadcasts, Beacons.
• Boosted device‑fingerprint entropy 20% with 450 browser feature tests.
• Implemented client‑side feature mapping for data privacy.
• Refined co‑visitation and other spam signals.
• Saved 30% in storage costs and reduced stream time 40% by shifting to S3Select.
• Sped‑up data analysis 64x using precompute and parallel stochastic algorithms.
• Built a revenue testing framework to measure spam signal quality.

Jan 2017 – Jun 2017 · 5 mos

• Constructed a scalable recommender system using Neo4j and Python, adopted by the Location Intelligence marketing team.
• Improved top-10 recall by 50% with preprocessing, kNN-Collaborative filtering.
• Created user lookalikes; Processed terabytes of GPS data with BigQuery

May 2016 – Jun 2016 · 1 mo · Bangalore

• Worked with the Anti-spam team.
• Performed bot detection using a combination of deterministic checks on headless browsers like phantom and selenium.
• Built a fully functional API using nodeJS and mongoDB. Spam was flagged on the basis of click frequency. Tested for the best fit model to calculate click threshold on the fly based on the user and website profile.

May 2015 – Jul 2015 · 2 mos · New Delhi, Delhi, India

• Implemented a Parallelized version of AKS algorithm using MPI.
• Additionally worked on RSA Encryption, primality testing and created my own Big Integer Library.