Aug 2024 – Sep 2025 · 1 yr 1 mo · Bengaluru, Karnataka, India · Hybrid

• ZDX Copilot (RAG + LLM Guardrails lead)
• Led design and development of RAG solution powering ZDX copilot using Knowledge graph, FAQ data source and playbooks
• Design and development of an automated QA framework to evaluate quality of answers generated by copilot
• Improve copilot guardrails to ensure safe use for 70+ design partners

Jul 2022 – Sep 2024 · 2 yrs 2 mos · Bengaluru, Karnataka, India · Hybrid

• AIOps project (Project lead)
• Led a team of Data Scientists and Data engineers towards the design, development, and production of the AIOps project to detect issues in zscaler datacenters (Time Series Forecasting, Anomaly Detection, LLM development)
• Help achieve precision of 97% with low alert noise, early detection of an average of 5 customer-impacting incidents per month
• ZPA Policy Recommendation
• Contributed towards core developments to improve the policy recommendation engine to reduce violations by 23% (Error analysis, EDA, Supervised, unsupervised learning algorithms - Kmeans, HDBSCAN, Random Forests)
• Proactively identified and addressed client concerns to drive product adoption and increase the number of recommendations accepted by 34%

Jul 2023 – Oct 2023 · 3 mos · Bengaluru, Karnataka, India · Remote

Oct 2019 – Jul 2022 · 2 yrs 9 mos · Southampton, United Kingdom

• Built and optimized the machine learning and deep learning models used in production to
• improve mean squared error by over 10 percent
• Created a framework that enables efficient hyperparameter search and tuning to generate a set of hyperparameters that helped achieve better performance
• Created a framework that can automatically search the best possible combination of
• models and their respective weights that maximize performance
• Helped solve the problem of over-prediction of sales on certain product lines by
• incorporating quantile regression into the solution that helped provide an estimate of
• the upper limit of sales
• Worked on implementing unsupervised learning algorithms to cluster sales trends of various
• department-category combinations and use the cluster information as a feature to
• encode the dept-cat information for the machine learning models
• Worked on research and development into explainability of the machine learning models and gain insights from them so that the end-users can understand, trust and implement the
• solutions in their businesses
• Built an optimization solution that helps retailers choose from a wide range of possible prices to optimize for key metrics and also set different constraints

Jul 2017 – Aug 2018 · 1 yr 1 mo · Bengaluru Area, India

• Developed and maintained data pipelines to support the generation of daily CXO-level scorecards, ensuring timely and accurate reporting.
• Designed and implemented an automation tool for quality checks, identifying and addressing potential issues in scorecards, saving the team an average of 2 hours per week and improving data reliability. I was awarded a "Spot Award" for this initiative.

May 2016 – Jul 2016 · 2 mos · Kolkata Area, India

• Worked on a full stack wedding planner application for the web
• Working for a startup helped me become accountable and deliver projects on a tight
• deadline