HARSH KUMAR
Data Science @ Uber, Gojek and American Express
MBA (IIM Indore), B.Tech (IIT Guwahati)
BIO
I have spent the last 11 years building data-driven solutions, robust experimentation strategies, and complex machine learning models for global tech companies like Uber, Gojek, and American Express. I design marketplace architectures, lead high-performing teams, and pioneer the adoption of AI coding agents with a focus on translating complex technical systems into massive business growth. I partner with global decision-makers to turn data into a distinct competitive advantage.
MY SKILLS
DATA & SYSTEMS
ENGINEERING
Languages: Python, SQL, Go, SAS
Data Architecture: BigQuery, Hive, PySpark
Cloud & MLOps: GCP, MLFlow, dbt, Git
MACHINE LEARNING
& GENAI
Core Modeling: XGBoost, LightGBM, Random Forest
Frameworks: PyTorch, TensorFlow, Scikit-Learn
Intelligent Systems: LLM Agents, Llama-Index, Deep Learning
EXPERIMENTATION
& MARKET STRATEGY
Methodologies: Advanced A/B Testing, Switchback Experiments
Causal Frameworks: Synthetic Controls, Causal Inference
Market Design: Pricing Strategy, Marketplace Liquidity, Network Dynamics
PROFESSIONAL EXPERIENCE
Led and executed the overarching data science initiatives for a pivotal business model shift from per-trip commissions to driver subscriptions across key growth markets.
- Marketplace Growth & Launch: Spearheaded the launch and experimentation strategy for driver subscription products in India, accelerating active driver supply by 15%+ and driving cross-LOB marketplace improvements.
- Pricing Frameworks: Designed the foundational pricing methodology for subscription products, balancing adoption, driver earnings, and marketplace liquidity at product launch.
- Advanced Experimentation: Established the experimentation strategy for the newly created subscriptions team, deploying over 10 randomized A/B tests and synthetic control experiments to isolate the impact of marketplace interventions.
- GenAI Evangelism: Acted as an early adopter and team-wide evangelist for AI coding agents (Claude Code, Cursor) through technical workshops, kick-starting the cycle of scaling data science productivity.
- Internal AI Tooling: Architected and deployed internal LLM agents to standardize the creation and peer-review of experimentation plans, drastically reducing preparation time while enhancing statistical rigor and consistency across global teams.
Data Science Manager
Data Scientist L5
Data Scientist L4
Data Analyst L4
Led machine learning, pricing, and experimentation initiatives across Ads, Pricing, and Marketplace Supply teams, driving platform reliability, revenue growth, and operational efficiency for Gojek's core marketplace.
- Pricing Strategy & Optimization: Spearheaded the ML pricing strategy for the Singapore market, delivering ~$50K in monthly gross take uplift by architecting a new optimization objective that balanced user growth with marketplace efficiency.
- Core Marketplace Systems: Directed the design and development of a next-generation driver positioning system and a Go-based order pooling module, improving recommendation quality by 15-30% while optimizing vehicle utilization.
- MLOps & Data Platform: Engineered and operationalized an end-to-end MLOps platform spanning ETL, automated training, and real-time model health alerting, reducing ad-hoc data infrastructure bandwidth by 66%.
- Advanced Experimentation: Designed robust validation frameworks, executing complex randomized A/B and switchback experiments to isolate network effects and measure the impact of high-stakes product rollouts.
- Architecture & Governance: Served as the sole Data Science representative in the Technology Architecture group, shaping engineering best practices and authoring/reviewing 6 Architecture Decision Records (ADRs/RFCs).
Senior Manager
Manager
Assistant Manager
Business Analyst
Led advanced analytics and machine learning initiatives supporting global Merchant Marketing and Services organizations, delivering predictive modeling, strategic consulting, and innovation platforms used for large-scale decision-making.
- Team Leadership & Mentorship: Managed and mentored a team of 4 data scientists and analysts, owning hiring, project prioritization, and global stakeholder engagement across multiple annual review cycles.
- Predictive Modeling Portfolio: Developed a portfolio of 15+ machine learning models (using XGBoost, GBM, KNN, and sequence mining) to predict customer life events and airline travel propensity.
- Risk & Operational Systems: Architected and deployed a global Merchant Anomaly Prediction system that detected terminal disruptions, customer suppression, and surcharge risks, significantly reducing merchant servicing costs.
- Analytics Innovation & Big Data: Engineered scalable analytics tools, including an Automated Insights Engine to eliminate manual data profiling, and a spatial Customer Location Tracker using DBSCAN and Hadoop MapReduce to infer consumer travel behavior.
- Strategic Consulting & Global Honors: Delivered data-backed site-selection and lifecycle re-engagement consulting to elite global merchants; awarded the American Express GCSG President's Award (awarded to the top 2-3% of performers globally).
- Designed & implemented highly scalable & flexible test harness framework
- Optimized test harness firmware to run from a small memory, reducing the total testing time
- Studied the build process for firmware & investigated possibilities of reducing boot time
EDUCATION
- Ranked 4th in a competitive batch of 447 PGP students.
- Senior Member, Equit-I (Finance Club) & Coordinator of the Finance Lab.
- Core Member, Finance Vertical for the i5 Summit.
- Bachelor of Technology (B.Tech) in Electronics and Communication Engineering
- Promoted to Electronics & Communication Engineering from Mechanical Engineering on the basis of academic merit.
- Awarded the prestigious Merit-cum-Means scholarship, covering full tuition fees and a monthly stipend.
