DTA Science

Down to a Science is presented by Galang AI. The views expressed by guests are their own, and guest participation does not constitute endorsement of Galang AI or its products or services. Despite the shared surname, Brandon has no relation to Galang AI.

In Episode 5, Brandon Galang, GTM Engineer at Vercel and author of the Substack Leverage Loops, joins Kevin G. Libuit to talk about why being technical no longer depends on writing code. Drawing on a path from industrial engineering to product management at Intel, Amazon, and Wonder, Brandon argues that engineering is really the ability to keep iterating toward a solution.

Kevin and Brandon discuss how long-form writing and Amazon's press release FAQ shaped Brandon's thinking, why working well with agents takes something like "theory of mind," and why Brandon would rather overload an agent's context window than under-explain. They also debate whether "taste" is a useful idea in the age of AI, how to break out of the new mean when everyone can build, and how Brandon found a side door into Vercel when every job posting draws hundreds of applicants. They close on the hidden cost of shipping too many tools, how Brandon decides which AI tools are worth learning, and how universities might prepare students for an AI-shaped workforce.

Show Notes: https://dtascience.com/episodes/episode-5

Music licensed through Soundstripe. Code: 2ABVRZT155WBCFJX

Chapters

00:00 Engineers iterate toward solutions

00:31 Meet Brandon Galang

02:08 Why long-form writing

05:15 Amazon's press release FAQ

07:52 Writing with AI

10:02 Technical vs. non-technical

11:05 What it means to be an engineer

13:51 From product manager to working with agents

15:11 Theory of mind for agents

18:11 Memento and the narrow world of agents

21:18 Personal agents and persistent context

23:35 Choosing an AI platform

25:25 Taste vs. convergence

30:08 Breaking out of the mean

35:03 Finding the side door to Vercel

37:00 Getting a hiring manager's attention

43:49 AI rewards tinkerers and generalists

47:58 The cost of shipping tools

50:26 Keeping up: the video game controller model

53:22 Find the bottleneck

55:20 Preparing students for an AI workforce

1:00:57 Where to follow Brandon

Down to a Science is presented by Galang AI. The views expressed by guests are their own, and guest participation does not constitute endorsement of Galang AI or its products or services.In Episode 4, Dr. Sandeep Pawar, VP and General Manager of Life Sciences at Manifold, joins Kevin G. Libuit to break down how real-world data has evolved across three eras: from claims data, to clinical data from electronic health records, to today's multimodal data spanning genomics, imaging, and wearables. He explains why the infrastructure for sharing and analyzing that data hasn't kept pace.Kevin and Sandeep discuss trusted research environments and why the conversation is shifting to "trusted execution environments" for AI. They also cover bringing your own tools to the data, deterministic versus probabilistic AI in regulated research, and why so many AI pilots built on clean data fall apart at scale. They close on why frontier AI labs are moving into life sciences, the value of wearable data, and who should be trusted with sensitive patient data.Show Notes: https://dtascience.com/episodes/episode-4Music licensed through Soundstripe. Code: 2ABVRZT155WBCFJXChapters00:00 Trusted research environments00:31 Meet Dr. Sandeep Pawar02:00 Real-world data, defined04:59 Era one: claims data10:33 Era two: clinical data from EHRs14:17 How clinical data shapes drug development17:10 Why pharma wasn't ready for the data19:30 Era three: multimodal data22:04 Making sense of a mountain of data24:45 Trusted research environments, explained27:57 Bring your own tools29:24 From trusted research to trusted execution31:13 Deterministic vs. probabilistic AI35:45 Forward-deployed engineers and why AI pilots stall40:10 AI implementation beyond biotech44:27 Why frontier AI labs want into life sciences47:19 Follow the money: GLP-1s vs. AI labs49:04 What AI can't digitize: the wet lab51:12 Wearables and the value of data52:29 Can you trust frontier labs with your data?55:36 Where to follow Sandeep and Manifold

Down to a Science is presented by Galang AI. The views expressed by guests are their own, and guest participation does not constitute endorsement of Galang AI or its products or services.

In Episode 3, molecular biologist, biotech entrepreneur, and educator Dr. Tshaka Cunningham joins Kevin G. Libuit to explore how early exposure, mentorship, and hands-on training can help build the biotech workforce. He shares how his grandmother’s laboratory, his mother’s creativity, and his great-grandmother’s commitment to service shaped his life—and why seeing Black scientists in leadership helped him imagine his own path.

The conversation follows his journey through pharmaceutical research, academia, government, and biotech entrepreneurship, including his work developing molecular diagnostics for behavioral health at Polaris Genomics. Kevin and Dr. Cunningham discuss LIFT-VA, Future Kings, GEMS, and Athari BioSciences; creating opportunities for students and instructors; and teaching scientists to work in teams, communicate, and defend their ideas. They close with Dr. Cunningham’s perspective on AI in drug discovery and why he encourages scientists to stay “above the AI layer” as the creative force behind their work.

Show Notes: https://dtascience.com/episodes/episode-3

Music licensed through Soundstripe. Code: 2ABVRZT155WBCFJX

Chapters

00:00 The human spark behind scientific creativity

00:34 Meet Dr. Tshaka Cunningham

02:16 A childhood in his grandmother’s lab

06:35 Three generations of science, art, and service

10:09 Finding a path through curiosity and representation

14:11 Mentorship and Black pioneers in science

19:44 Choosing a scientific career

24:20 Learning drug discovery at Bristol Myers Squibb

25:37 How an idea becomes a medicine

29:50 From Rockefeller to government science

33:08 Connecting academia, industry, and government

38:19 From federal service to biotech entrepreneurship

41:56 Molecular diagnostics for behavioral health

44:50 Building the biotech workforce through LIFT-VA

49:09 Industry partnerships and training future leaders

51:38 Expanding early-career scientists’ options

53:37 Teamwork, capstones, and communicating science

56:00 Preparing scientists for AI

59:15 Creativity and staying above the AI layer

1:03:53 AI’s potential in drug discovery

1:05:45 Closing reflections on service

Down to a Science is presented by Galang AI. The views expressed by guests are their own, and guest participation does not constitute endorsement of Galang AI or its products or services.In this episode, Dr. Luke Short, Director of the Dallas County Public Health Laboratory, joins Kevin G. Libuit to explore AI adoption in public health. From reviewing standard operating procedures to building spreadsheets and scripts that can be inspected and validated, Luke shares practical ways AI can support laboratory teams while keeping human judgment and accountability at the center.We discuss his journey from chemistry to public health leadership, privacy in regulated environments, workforce capacity, and why learning to use new technology can start with something as simple as cooking. We also explore Jevons paradox, the changing demand for software, and Luke’s nuanced take on the “iPad generation.”Show Notes: https://dtascience.com/episodes/episode-2Music licensed through Soundstripe.Code: 2ABVRZT155WBCFJX

Down to a Science is presented by Galang AI. The views expressed by guests are their own, and guest participation does not constitute endorsement of Galang AI or its products or services.

In this episode, David Aanensen shares his journey from neuroscience and web development to leading innovations in pathogen genomics and public health surveillance. We explore how technology, visualization, and AI are transforming infectious disease monitoring and the importance of data sovereignty.