Shabbir Governor
Summary
Data Science professional with 8+ years of experience across AdTech, Fintech, and Insurtech. Specialised in building production-grade machine learning systems — from real-time bidding engines and two-tower recommendation models to Generative AI products and full-stack applications. Passionate about converting complex data into scalable, high-impact solutions.
Professional Experience
Lead Data Scientist
Mobavenue Media · April 2025 – Present
Real-Time Bidding · Two-Tower Models
Architected a production Two-Tower deep learning system for real-time bidding — a User/Device Tower paired with an Ad/Creative Tower, joined by learned dot-product similarity scoring to predict bid prices across CPI, CPA, and ROAS objectives at massive scale.
- ▸Trained separate CPI, CPA, and ROAS two-tower models, each ingesting TBs of click, install, and post-install event data daily for objective-specific optimisation.
- ▸Deployed on a horizontally scalable inference cluster serving 10B+ bid requests/day across global RTB exchanges with sub-80ms response times.
- ▸Built online learning pipelines for continuous model refresh, keeping embeddings in sync with shifting user behaviour without costly full retrains.
- ▸Replaced heuristic bid multipliers with learned per-auction probability scores, delivering measurable ROAS uplift across $50M+ in managed ad spend.
- ▸Designed dynamic eCPM prediction and budget pacing algorithms for smooth, efficient campaign delivery across supply sources.
- ▸Established real-time data pipelines (Kafka + PySpark) feeding both model training and live operations dashboards.
Lead Generative AI Engineer
Angel Broking · April 2024 – March 2025
- ▸Built a Creative Design Tool optimising ₹7 crore+ in ad spend, improving targeting and campaign efficiency.
- ▸Developed RAG pipelines for document extraction, summarisation, and cross-departmental Q&A.
- ▸Engineered AI-driven stock and mutual fund recommendation systems using advanced LLMs and prompt engineering.
- ▸Automated NLP-powered document workflows, reducing manual processing by 30%.
Senior Data Scientist
BLKBOX.AI · April 2021 – March 2024
- ▸Designed an ad-generation pipeline using transfer learning and semantic similarity, deployed on AWS.
- ▸Used Prophet time-series forecasting to predict ad performance across Facebook, Google Ads, and Twitter in real-time.
- ▸Built LSTM and GAN-based models to predict creative asset performance and recommend similar live campaign ads.
- ▸Created and maintained Tableau dashboards for dynamic business analytics.
Professional Product Developer – DS
DXC Technologies · November 2018 – February 2020
- ▸Built predictive models on IBM i DB2 insurance data to assess customer behaviour and claims patterns.
- ▸Designed churn prediction models for Nippon Life Insurance, improving retention through real-time predictions.
- ▸Applied R-based anomaly detection to identify irregularities in claims data.
- ▸Built Tableau dashboards for new insurance product launches.
Key Skills
Languages
ML & AI
Technologies
Education
PGP in Data Science & Big Data (2020–2021)
Aegis School of Data Science, in association with IBM · Mumbai
Bachelor of Engineering (2014–2018)
Mumbai University, SJCET · 7.77 CGPA