Leading initiatives to enhance our AI/ML model training platform:

1. Building enterprise LLM training (pre-training, fine-tuning, distillation, post-training) infrastructure with SFT/RL pipelines

2. Managing training platform supporting core ML workflows, driving migration from legacy systems to modern ML automation platforms for seamless model deployment cycles

3. Enhancing modeler/ML developer experience through AI tooling and agentic automation, to reduce model time-to-market

4. AI Compute: Building seamless auto scalable compute platform with optimizations at application and infrastructure layers to improve overall GPU utilization

For the inaugural Women's Day Conference: Thrive with Vahati (https://thrivewithvahati.org/), I got to:

1. Design the 2-day Program end to end - Planning Day 1 and Day 2 of conference, Coordinating Speakers and Panelists, Identifying the right questions to be asked

2. Bring in largest Corporate Sponsorship to date

3. Launch the BAY AREA Chapter!

Started as the Machine Learning PM for one product and soon scaled to manage both the Product and Platform Data Science portfolio across all FSQ products. Work involved:

1. Product DS: Building ML models and algorithms to improve Places and Users data quality

• Places: accurately predicting open/closed POIs, choosing lat-long for a place from 100+ sources, harmonizing POIs based on core attributes, photo classification

• Users: understanding how people move and transact across places, normalizing foot traffic data to scale to real world trends, identifying demographics like age, gender, income using movement and transaction data

2. Platform DS: Done in 2 parts,

• Establishing standard procedures for ML lifecycle from annotation collection, setting custom release criteria, impact measurement, model serving to continuous model improvements and retraining

• Setting up required ML infrastructure like feature store, monitoring for improving the efficiency of model development and maintenance

Joining when Ally was an early stage startup, I had an interesting time deeply understanding what problems customers faced qualitatively & quantitatively, what Engagement & Retention metrics really makes sense for us and driving initiatives to improve them as we scale. Key Projects:

1. Created & Institutionalized Product Engagement & Retention metrics across Customer Lifecycle

2. Established a deep and robust Voice of Customer program and NPS engine combining qualitative research & quantitative data to continuously inform Product roadmap

3. Built data pipeline for Growth Flywheel and a cohesive Customer 360 profile to establish a single source of truth

1. Led Product Development for an instant end-to-end User Engagement experience, infinitely scalable to 100M+ users:

• Set up a streaming big data platform for real time analytics with Apache Kafka, Spark and HDFS

• Built real time fulfillment engine with multiple Incentive Types, Omni channel communication & comprehensive Campaign Database

2. Built 0-1 Loyalty Product for Jio Financial Services to improve product adoption & retention

3. Portfolio Management: Led a team to run targeted campaigns; Observed lift >20%;