Jul 2023 – Jul 2025 · 2 yrs · San Francisco Bay Area

• Spearheading an AI-driven macro research initiative using deep learning and transformer models to forecast complex, real-world time series.
• Applied transfer learning and LLM-based sentiment analysis to extract insights from unstructured text and noisy financial data.
• Collaborated with cross-functional teams to deploy robust, scalable PyTorch pipelines using Airflow, balancing speed of iteration with production reliability.
• Focused on delivering research-to-production impact in high-stakes, data-constrained environments.

Oct 2022 – Dec 2022 · 2 mos · Stanford, California, United States · On-site

• Defined strategic vision and go-to-market plans for scalable, interactive AR applications in NBA stadiums, identifying $12M+ revenue potential.
• Engaged stakeholders across sports, telecom, and media sectors to validate use cases and ensure product-market fit.
• Developed an MVP leveraging Niantic’s platform to showcase innovative AI and AR integration for real-world fan experiences.
• Explored AI-driven product management through Stanford’s EE205 course, collaborating closely with Niantic on AR fan engagement experiences.

Jun 2022 – Sep 2022 · 3 mos · Mountain View, California, United States

• Designed and implemented a BERT-based multi-class webpage classification model to optimize enterprise search relevance and reduce maintenance overhead by 70%.
• Engineered a hybrid data pipeline integrating web crawling, URL heuristics, and NLP embedding-based clustering to handle noisy and imbalanced datasets.
• Applied advanced techniques like label smoothing and human-in-the-loop annotation workflows to mitigate label noise and improve model robustness.
• Scaled the system to support onboarding and accurate classification of 100+ enterprise business applications, enabling seamless integration and search optimization.

Jun 2019 – Jul 2021 · 2 yrs 1 mo · Gurugram, Haryana, India

• Led development of machine learning strategies for alpha generation, leveraging LSTMs, random forests, and genetic algorithms to navigate noisy, non-stationary financial time series.
• Designed research workflows that balanced empirical rigor with rapid iteration—optimized generalization across volatile market regimes.
• Built scalable, low-latency infrastructure in Python and C++ for data ingestion, feature generation, model training, and backtesting, supporting production deployment with millisecond-level latency constraints.
• Deployed real-time monitoring and diagnostics systems to detect model drift, infra regressions, and changing data patterns—ensuring robust performance in production.
• Contributed to a culture of iterative experimentation, automation, and metrics-driven evaluation within a high-performance research environment.

May 2018 – Jul 2018 · 2 mos · Gurgaon, India

May 2017 – Jul 2017 · 2 mos · Austria

Dec 2016 – Dec 2016 · 0 mo · Bengaluru Area, India