Building AI That Augments, Not Replaces People in Logistics with Shaman Ahuja, Deputy CEO of Optym

We explore why multi-objective optimization matters more than finding a single "right" answer, how Optym balances traditional algorithms with modern AI, and why garbage in, garbage out is still the biggest challenge in optimization. Shaman also shares hard-won lessons about organizational design, including why platform teams often fail and how Optym structures product teams around business problems instead.

You'll hear about the surprising parallels between airline operations and freight, why planners with 20+ years of experience won't trust black box solutions, and how LLMs can serve as an interface layer without replacing the heavy lifting of discrete optimization models.

Key Topics

✦ Wall Street to freight tech: career transitions in logistics

✦ Augmentation vs. automation in AI implementation

✦ Multi-objective optimization across airlines, LTL, and truckload

✦ Why business-focused teams beat platform teams

✦ Building enterprise software with a customer-first mindset

✦ The role of LLMs alongside traditional optimization algorithms

✦ Lessons from 25 years of bootstrapped growth

Segments

00:00 - Intro

01:06 - What is Optym? Transportation Optimization Explained

03:26 - Shaman's Background

05:51 - Goldman Sachs Experience: Life on Wall Street

07:59 - Business Beyond Money

09:14 - Parallels Between Finance and Logistics Optimization

11:25 - Customer Focus: From Southwest Airlines to LTL

13:33 - Time to Value: Delivering on Promises Consistently

15:28 - Complexity Across Transportation Modes

16:25 - The Faster Horse Problem: Balancing User Feedback with Innovation

18:15 - Multi-Objective Optimization: No Single "Right" Answer

21:12 - Balancing Traditional Algorithms with Modern AI

23:18 - Garbage In, Garbage Out: Data Quality Matters

25:49 - LLMs as Interface Layer, Not Replacement

26:00 - Why Black Box AI Doesn't Work

27:21 - Airline Network Planning: 9 Months Ahead Scheduling

29:05 - Matching Capacity to Demand: Airlines vs Trucks

30:28 - LTL Dock Optimization: Learning from Airlines

34:18 - U-Haul Pod Optimization Case Study

36:17 - Organizational Design: Platform Teams vs Business-Focused Teams

39:10 - The 80/20 Split: Dual Hats in Product Development

40:45 - Building With Care: Why Deep Domain Knowledge Takes Years

44:34 - 25 Years Bootstrapped: Growth Without VC Funding

46:16 - Multi-Objective Thinking: AI as Augmentation, Not Replacement

48:30 - Final Thoughts: Context, Culture, and Getting AI Right

Whether you're building technology for transportation or implementing AI in operations, this conversation offers a refreshing perspective on doing more with less—without losing the human expertise that makes it all work.

Show Information

This Side Up Podcast digs deep into the minds of founders, builders, and operators reshaping the future of commerce and logistics. Subscribe for more conversations with industry leaders who are actually building the future.

Podcast Website: https://www.thissideup.fm/

Produced by: Intermode - https://www.intermode.is/

Host: Andrew Verboncouer - https://www.linkedin.com/in/andrewverboncouer/