Jun 2025 – Dec 2025 · 6 mos · Menlo Park, CA · Hybrid

• Instagram Home Relevance - ML Foundation [Model Fundamentals]
• Working as part of the Instagram Home Relevance team, I focused on ensuring the stability, reliability, and performance of large-scale machine learning systems that power content ranking and recommendations for billions of users. My work involved close collaboration with engineers, product partners, and data scientists to monitor model health, investigate anomalies, and connect low-level model signals to top-line product outcomes. My responsibilities included:
• 1) Model Health Monitoring & SEV Response: Led ongoing monitoring for Instagram ranking models, identifying anomalies and severity events early, reducing SEV detection time and protecting experience quality for 3B+ monthly active users.
• 2) Metric Deep Dives & Root Cause Analysis: Conducted deep dives into engagement and ranking metrics to explain shifts in performance, uncovering underlying drivers and translating findings into clear, actionable insights.
• 3) Dashboarding & Signal Alignment: Reviewed and improved unified dashboards linking model-level signals to product metrics, cutting investigation time and supporting decision-making for 50+ cross-functional stakeholders.
• 4) Product Impact & Recommendations: Partnered with product and engineering teams to translate analytical findings into recommendations that improved content relevance and feed quality.
• These efforts helped strengthen the reliability of Instagram’s recommendation systems, improve observability across ML pipelines, and ensure model behavior aligned with user experience and business goals.

Aug 2024 – Feb 2025 · 6 mos · San Diego, California, United States · Hybrid

• In this role, I contributed to enhancing transportation modeling through data analysis, summarization, and process improvements. My responsibilities included:
• 1) Meta-Data Management: Adding and managing metadata for ABM3 tables in the datalake to enhance data discoverability and standardization.
• 2) Model Output Analysis & Dashboard Development: Working with model outputs to create Power BI dashboards, enabling easier interpretation for urban planners and stakeholders.
• 3) Automated QC Development: Implementing quality control automation in Databricks for post-ETL processes, leveraging machine learning techniques to improve data integrity.
• 4) Model Output Summarization & Comparison: Documenting and summarizing model outputs, comparing older and newer EMME model versions to identify key differences and performance improvements.
• 5) Documentation: Creating and maintaining thorough documentation for processes and models, ensuring consistency and clarity across the team.
• These contributions helped optimize data workflows and enhanced the accuracy of transportation modeling outputs for regional planning.

Jun 2024 – May 2025 · 11 mos · San Diego, California, United States · Hybrid

• Working in collaboration with the Cosmology group at UC Berkeley and UC San Diego, I focused on analyzing and optimizing astronomical observational datasets to enhance the accuracy of astrophysical research. My work involved:
• 1) Data Exploration & Parameter Analysis: Initiated with an in-depth understanding of the PB2b dataset, identifying key parameters such as azimuth, elevation, bolostage temperature, and CHWP frequency. Conducted trend analysis to identify patterns and relationships, leveraging tools like SQL and JSON extraction.
• 2) Automating Data Pipelines: Developed automated scripts to streamline the data extraction and analysis process, improving efficiency in parameter monitoring and reporting.
• 3) Database Management:Enhanced existing science analysis databases to support ation for comprehensive analysis.
• These efforts contributed to improving the precision of astrophysical measurements and advancing our understanding of cosmic phenomena.

Aug 2023 – Oct 2023 · 2 mos · San Diego, California, United States · Remote

• Working in collaboration with J-PAL Lab at MIT, I leveraged advanced analytics and tools to significantly enhance environmental assessments, supporting strategic decision-making and operational efficiency in environmental research projects.
• Enhancing environmental assessments is critical for accurate and efficient research, which in turn supports informed policy-making and resource management. By automating web scraping and data processing, I saved 20 hours weekly, increased data availability by 50%, and improved data accuracy by 30%. These efforts streamlined operations, reduced manual workload, and provided reliable data for critical environmental decisions.
• Achievements:
• 1) Web scraping process automation leading to saving of 20 hours per week
• 2) Boosted data accuracy for vegetation data with Google Earth Engine
• 3) Enhancing data analysis efficiency transforming unstructured data to gain meaningful insights
• These advancements were crucial in advancing our environmental research capabilities, enabling more accurate and efficient analysis.

Jun 2023 – Aug 2023 · 2 mos · San Francisco Bay Area · Remote

• Customer retention is crucial for business sustainability and growth. Retaining customers reduces the cost of acquisition, increases lifetime value, fosters brand loyalty. My work at 6sense directly contributed to these goals by using data-driven insights to predict and mitigate churn, ensuring ongoing customer satisfaction and loyalty.
• Achievements:
• 1) Significantly reduced attrition rates by 85% and identified key factors for targeted retention efforts, bolstering the company's strategic objectives
• 2) Enhanced data analysis capabilities, leading to more informed strategic decisions
• 3) I developed a predictive model that decreased customer churn by 25% and improved decision-making effectiveness by 15%
• This role taught me the importance of data-driven decision-making and enhanced my ability to turn complex data into strategic insights.

Jan 2022 – Jun 2022 · 5 mos · Gurugram, Haryana, India · Remote

• During my tenure as a Data Science Engineer at BlackRock, I specialized in automating and optimizing data processes using Java, Apache Ignite, Microsoft Azure, and Agile SDLC methodologies.
• By streamlining data processes, I enhanced operational efficiency and reduced manual workload, crucial for maintaining BlackRock's competitive edge. Automating data tasks and improving system reliability directly impacted the company's ability to manage large-scale financial data efficiently.
• Achievements:
• 1) Developed a Java application to automate end-to-end testing of BlackRock's GPExceptionServer, cutting manual error-checking time by 50% and increasing scalability for multiple testing scenarios
• 2) Enhanced data integration within BlackRock's Aladdin platform by automating CSV to map conversions, significantly reducing manual intervention
• 3) Improved system reliability and efficiency by 40%, advancing BlackRock’s operational capabilities
• These efforts were instrumental in advancing BlackRock's data management processes and operational efficiency.