Sci-Fi in the Wet Lab: Building an Agentic, Self-Driving Cell Culture Lab with t

What happens when generative AI and data science step off the screen and directly into the physical wet lab?

In this talk and live demonstration, builder Dale Herzog reveals how open-source lab automation, modern generative AI, and data science pipelines are turning autonomous biology from science fiction into accessible reality. Leveraging a unique multidisciplinary background spanning mechanical CAD design, industrial automation, data science, and generative AI engineering, Dale demonstrates an end-to-end, closed-loop cell culture platform built around a modified Opentrons OT-1 liquid handling robot.

Attendees will see how the entire experiment lifecycle is automated and orchestrated without human intervention:

Agentic Generative AI Orchestration: LLM-driven agents ingest experimental hypotheses, generate executable liquid-handling protocols, and dynamically troubleshoot edge cases mid-run.

Multimodal Vision Feedback (VLMs): Vision-Language Models inspect microscopic assay imagery and assess cell culture morphology and viability in real time, serving as the automated "eyes" of the lab.

Data Science & Closed-Loop Analysis: Assay data and sensor telemetries are parsed, structured, and fed directly back into reasoning models to decide subsequent dosing, incubation, or harvesting parameters.

Digital Fabrication & Physical AI: Custom SolidWorks-engineered 3D hardware—such as PETG flask shakers, automated decappers, and motorized microscope gantries—bridges the gap between off-the-shelf laboratory robotics and custom biological workflows.

Dale will share practical insights on how he used generative AI to accelerate embedded firmware calibration and driver synthesis, how data-centric pipelines keep experimental iterations reproducible, and how open-source fabrication enables individuals to build self-driving labs at a fraction of traditional pharmaceutical automation costs.

Key Takeaways & Highlights:

Closed-Loop Scientific Discovery: How agentic LLMs translate scientific objectives into physical commands and use real-time VLM vision feedback to evaluate results.

Data Science in the Wet Lab: Architecting data pipelines to handle automated imaging, assay metrics, and time-series telemetry for closed-loop decision making.

Full-Stack Hardware & AI Prototyping: Practical lessons from designing custom benchtop instruments in SolidWorks, 3D printing peripherals, and writing hardware drivers with GenAI coding assistants.

The Democratization of Biotech: How combining open robotics (Opentrons) with modern AI toolchains lowers the barrier of entry for automated biology and viability testing.

About the Speaker:
Dale Herzog is a mechanical CAD designer, data science practitioner, and generative AI builder specializing in industrial automation, digital fabrication, and open-source lab instrumentation. By bridging physical engineering (SolidWorks, rapid prototyping, robotics) with software intelligence (data science pipelines, agentic orchestration, and multimodal LLMs), Dale focuses on creating modular, low-cost self-driving laboratory systems that democratize scientific experimentation.

Agenda

---

Hosted By

Csaba Toth, GDG lead, WTM ambassador

Generative AI Engineer, GDG lead, WTM ambassador, tech meetup enthusiast.

Jacob Tonna, team member

Xayaphone Salinthone, Co-lead

I am passionate about helping others through technology, whether in the workplace or within my community, and have consistently been the person people turn to for troubleshooting, guidance, and support. I bring a strong commitment to continuous learning and I am driven to expand my skills in IT systems, networking, and emerging technologies while contributing to efficient, user-focused solutions.

Complete your event RSVP here: https://gdg.community.dev/events/details/google-gdg-fresno-presents-sci-fi-in-the-wet-lab-building-an-agentic-self-driving-cell-culture-lab-with-the-opentrons-ot-1/.