Devakinandan Panda
Staff II Software Engineer at Coupang | Distributed Systems, Logistics & Applied AI | ex-Amazon, ex-Roblox
Seattle, Washington, United States
Actions
Devakinandan Panda is a Staff II Software Engineer at Coupang's Operations Technology organization, with over 15 years designing and leading large-scale distributed systems across e-commerce, logistics, and gaming at Amazon, Roblox, and Coupang.
At Amazon he was the founding engineer of the automated Delivery Routing and Assignment platform, whose routing and assignment algorithm was deployed across all Amazon Grocery and Amazon Fresh delivery stations and later adopted for Quick Commerce ultra-fast delivery. He also led Amazon's company-wide migration to NoSQL databases. At Roblox he led the Avatars Platform, migrating 30 billion records live to CockroachDB with zero major incidents. At Coupang he architects an end-to-end logistics simulation digital twin and led organization-wide adoption of generative AI and LLM tooling, scaling from 5% to 100% of the engineering organization in under a quarter.
He is an IEEE and ACM member, was named IAOTP Top Engineer of the Year in Software Development in 2025, and speaks on distributed systems, routing and dispatch optimization, applied AI for logistics, and measuring engineering velocity with DORA metrics.
Area of Expertise
Topics
Measuring What Matters: Using DORA Metrics to Prove Your AI Dev Tools Actually Work
Every vendor promises their tool will make your team faster. Almost none can prove it — and neither can most teams that adopt them. This talk shows how to replace "it feels faster" with evidence, using DORA metrics (deployment frequency, lead time for changes, change-failure rate, and mean time to recovery) as an objective baseline for evaluating developer tools and AI assistants.
Drawing on establishing DORA baselines across an engineering organization to quantify the impact of new tooling and drive investment decisions, we'll cover how to instrument the four key metrics without a six-month project, how to run a before/after comparison that survives scrutiny, the traps that produce misleading numbers, and how to turn results into a clear decision: double down or cut the tool.
Attendees leave able to build a measurement framework that tells them, with data, which of their tools are actually worth the money. Vendor-neutral and practical, for engineering leaders, DevOps and platform engineers, and anyone accountable for tooling and AI ROI.
From 5% to 100% in One Quarter: Driving Real GenAI Adoption Across an Engineering Org
Most companies have bought AI coding tools. Far fewer have gotten their engineers to actually use them, and fewer still can prove it moved the needle. This talk is a practical, field-tested playbook for taking generative-AI and LLM tooling from a handful of early adopters to full-team adoption, drawn from scaling GenAI usage from 5% to 100% across an engineering organization in under a single quarter.
We'll cover the real blockers (trust, workflow fit, and 'I tried it once' fatigue), the rollout tactics that actually work versus the ones that stall, how to set guardrails without killing momentum, and, critically, how to measure genuine productivity gains rather than vanity usage numbers. Attendees leave with a concrete adoption roadmap they can run inside their own org, whether they lead a team of five or an organization of hundreds.
Aimed at engineering leaders, platform and developer-experience teams, and anyone accountable for making an AI investment actually pay off.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top