Super Data Science Podcast: Matt Glickman on AI Agents, Data Engineering & AI-First Enterprises

AI agents are changing enterprise data engineering — and February 2026 may be remembered as the “event horizon.”
In this episode of the Super Data Science Podcast, Matt Glickman, Founder & CEO of Genesis Computing, joins the show to discuss why enterprise AI has crossed a major threshold — and why data engineering is one of the highest-leverage places to apply agentic AI.

Matt shares lessons from nearly 25 years at Goldman Sachs, his move to Snowflake during the early cloud data wave, and why he now believes AI agents can help enterprises run faster, leaner, and smarter. The conversation covers the hidden knowledge inside large organizations, why data engineers are overloaded, how Genesis uses agents, blueprints, missions, and living context graphs, and why enterprise AI must prioritize correctness, auditability, and proof of work over novelty.

You’ll learn
- Why Matt calls February 2026 an AI event horizon for computing and enterprise work
- How Goldman Sachs’ data platform challenges foreshadowed the Snowflake opportunity
- Why financial services and healthcare are early AI adopters despite being late to cloud
- Why data engineering is painful, underappreciated, and full of hidden institutional knowledge
- How Genesis Computing automates data engineering with AI agents inside the customer’s environment
- What living context graphs are and how they preserve enterprise knowledge before it walks out the door
- Why agents need confidence checks, escalation, artifacts, tests, and proof
- Why enterprise AI needs correctness more than creativity
- How AI changes hiring, junior roles, education, and the future of data work
- How to know when a technology shift is real enough to stop watching and start building

Who it’s for: data engineers, data scientists, AI leaders, enterprise architects, founders, investors, CIOs, CTOs, and anyone trying to understand how agentic AI will reshape enterprise data work.

⏱️ Chapters
00:00 Intro: February 2026 as an AI event horizon
00:52 Welcome: Matt Glickman joins Super Data Science
01:40 Goldman Sachs, the financial crisis, and the power of data platforms
03:29 The “big user problem” before cloud data platforms
04:18 Early Snowflake meeting at Goldman Sachs
06:08 Why Matt left Goldman to join Snowflake
08:21 Building Snowflake bi-coastally: Bay Area engineering, New York customers
09:42 Why New York is uniquely strong for enterprise AI
11:39 Finance and healthcare as early AI adopters
13:26 February 2026 and the leap in model capabilities
14:49 Why institutional knowledge is the missing context layer
15:21 Data engineering burnout and knowledge walking out the door
16:43 Context graphs: reading databases, code, emails, docs, and workflows
19:16 Genesis Computing and the mission to create AI-first companies
20:37 Why enterprises must ask: “Why can’t an agent do this?”
21:38 Why Genesis focuses on the data engineering bottleneck
25:21 Genesis as an agent platform for data engineering
26:28 AI agents testing 30,000 reports instead of sampling
27:21 Error rates, proof of work, and why humans make mistakes too
28:34 Confidence checks, artifacts, and agent verification
31:09 Will AI replace data engineers? The junior hiring shift
33:41 The rise of AI conductors managing multiple agents
35:40 Why schools need to teach AI fluency now
38:24 How customers onboard Genesis like a new employee
39:46 Example use case: asset management dashboard from raw feeds
40:36 Blueprints as guardrails for complex data workflows
42:34 Finance and healthcare adoption, secure deployment, and customer environments
44:32 Why agents should escalate when confidence is low
46:00 “Show me the artifact”: proving work before proceeding
48:38 Correctness vs. novelty in enterprise AI
50:57 Living context graphs and compounding institutional knowledge
56:06 Four-phase AI adoption: assessment to scaled autonomy
58:50 How to convince enterprises: show, don’t tell
01:01:51 How to know when a technology shift is real enough to build
01:06:43 Book recommendation: The Hitchhiker’s Guide to the Galaxy
01:10:01 Where to find Matt and Genesis Computing
01:11:03 Episode recap and closing notes

Key takeaways
- Enterprise AI is not just about better models — it requires context, guardrails, proof, and trusted workflows.
- Data engineering is a prime use case for agents because it combines business logic, code, data, documentation, and repetitive validation.
- The biggest long-term advantage may be turning tribal knowledge into a company-owned living context graph that compounds over time.

If this helped
👍 Like • 🔔 Subscribe • 💬 Comment

What enterprise workflow should AI agents automate first — data engineering, reporting, testing, or documentation?

#DataEngineering #AIAgents #AgenticAI #EnterpriseAI #LLMOps #Snowflake #FinancialServices #AIStartups #FutureOfWork #GenesisComputing

Видео Super Data Science Podcast: Matt Glickman on AI Agents, Data Engineering & AI-First Enterprises канала Genesis Computing