Speaking
I love helping audiences see the hidden science behind systems they interact with every day. Whether it's a keynote, a seminar, a podcast, or a guest lecture — I bring real stories from building optimization and AI systems at global scale, and I make complex ideas accessible without oversimplifying them.
Available formats: Keynote, Conference Talk, Seminar, Workshop, Panel, Podcast, Guest Lecture
Talk Topics
Each talk is grounded in real systems and real outcomes — not theory alone
Demand Forecasting at Scale
Your forecasting system works in the lab. But what happens when you need a million models running concurrently across six countries, handling 300% volume surges during peak? This talk covers the architectural, statistical, and organizational challenges of forecasting at massive scale.
For: Data science leaders, supply chain executives, ML engineers
Formats: Keynote, Seminar, Workshop
What your audience will take away:
- ▸ How to architect forecasting systems that scale from hundreds to millions of concurrent models
- ▸ Practical ensemble strategies combining classical statistics, deep learning, and tree-based methods
- ▸ Organizational patterns for managing forecasting across global regions and planning horizons
Last Mile Logistics Optimization
The last mile is the most expensive, most complex, and most visible part of the supply chain. This talk explores the science behind optimizing package delivery across 1,000+ stations — from carrier assignment to route densification to cost estimation — and why the hardest problems are often the ones customers never see.
For: Operations leaders, logistics executives, OR practitioners
Formats: Keynote, Conference Talk, Panel
What your audience will take away:
- ▸ How optimization and ML work together to reduce delivery costs by hundreds of millions annually
- ▸ Why granularity matters — moving from 80 zones to 50M+ geographic blocks transforms decision quality
- ▸ Real-world tradeoffs between speed, cost, and capacity in delivery networks
Bridging OR & ML in Production Systems
The most impactful industry systems don't choose between operations research and machine learning — they combine both. This talk draws on real examples to show how to design architectures that leverage optimization and learning together, and why the gap between these disciplines costs companies billions.
For: Technical leaders, research scientists, engineering managers
Formats: Keynote, Seminar, Guest Lecture
What your audience will take away:
- ▸ Design patterns for hybrid OR+ML systems that actually ship and scale
- ▸ When to use optimization vs. learning vs. both — a practical decision framework
- ▸ How to build teams and cultures that bridge the OR-ML divide
From Research to Billion-Dollar Systems
The journey from PhD research to deployed science generating hundreds of millions in value. What changes when you move from papers to production? What stays the same? And what do the textbooks never teach you about making science work in the real world?
For: PhD students, early-career researchers, academic collaborators
Formats: Guest Lecture, Seminar, Podcast
What your audience will take away:
- ▸ How to identify which academic research has real production potential — and which doesn't
- ▸ The skills that matter most when transitioning from academia to industry research
- ▸ How to measure and communicate the impact of deployed science
Capacity Planning Under Uncertainty
When demand is volatile and the cost of being wrong is enormous, how do you plan labor, vehicles, and infrastructure? This talk covers multi-stage optimization approaches to capacity planning, drawn from real systems handling millions of daily decisions.
For: OR practitioners, supply chain planners, capacity managers
Formats: Conference Talk, Seminar, Workshop
What your audience will take away:
- ▸ Practical approaches to stochastic capacity planning that balance cost, service, and resilience
- ▸ How to build planning systems that adapt gracefully to 300% demand surges
- ▸ Lessons from deploying multi-stage newsvendor optimization at scale
Speaker Bio
Copy the version that fits your program
Chinmoy Mohapatra is a Senior Research Scientist at Amazon, where he builds large-scale optimization and forecasting systems for last-mile logistics. His work spans demand planning, route optimization, and capacity allocation across six global regions. He holds a PhD from UT Austin and 10 USPTO patents.
Chinmoy Mohapatra is a Senior Research Scientist at Amazon, where he leads the development of optimization and forecasting systems for last-mile delivery at global scale. His systems manage over one million concurrent forecasting models and process millions of packages daily across Amazon's global delivery network. Previously, he served as Lead Operations Research Scientist at BNSF Railway, where his work on train routing optimization contributed to BNSF receiving the 2018 INFORMS Prize — the highest organizational honor in operations research. Chinmoy holds a PhD in Supply Chain and Operations Management from the University of Texas at Austin, has 10 USPTO patents, and has presented at 36+ conferences including INFORMS, POMS, and MSOM.
Speaker Assets
Speaker One-Sheet
PDF overview of topics, bio, and contact
Coming soonPast Appearances
Selected conferences, invited talks, and panels
POMS Annual Conference
Data Science, Analytics, and AI
INFORMS Annual Meeting
Network Analytics at Amazon Last Mile Logistics
POMS Annual Conference
Optimizing Demand and Capacity Planning in Last Mile Logistics
INFORMS Annual Meeting
Error-Based Forecast Refinement
INFORMS Annual Meeting
Amazon Last Mile Warehouse Operations
INFORMS Annual Meeting
Capacity and Coincidence Aware Package Assignment
MSOM Conference
Delivery Pricing Strategies in Retail Competition
INFORMS Annual Meeting
BNSF OR Tools Showcase & Block Re-design Optimization
Indian School of Business
Optimal Subscription Pricing for Free Delivery Services
MSOM Conference
Optimal Subscription Pricing for Free Delivery Services
Invite Me to Speak
I'm available for conferences, corporate events, university lectures, and podcasts. To help me prepare the best possible talk for your audience, please include:
- ▸ Event name and date
- ▸ Format (in-person, virtual, hybrid)
- ▸ Expected audience and size
- ▸ Topic of interest or preferred theme
Typical response time: 2-3 business days