Keywords: Robotics, Cell Culture Automation, Life Sciences, SLAS2026, Whelix, Lab Automation, AI, Autonomous Systems, ROS2, Jetson Orin, Spring Boot, Electron, React

Learn more at whelix.co

Overview

As Head of Robotic Engineer at Whelix, I lead the design, development, and deployment of the robotic software stack powering our flagship product. In February 2026, the Whelix team brought the Associan — our end-to-end fully autonomous cell culture robotic system — to SLAS 2026 (Society for Laboratory Automation and Screening Annual Conference), one of the premier stages for laboratory automation and life science innovation.

Over three days on the expo floor, the Associan attracted 1,000+ audience members and generated significant interest from industry leaders including Festo, ThermoFisher Scientific, and Boston Children’s Hospital. This post chronicles the engineering story behind the system, the logistics of getting it there, and the reception it received from the global life science community.

What is Whelix Associan?

The Whelix Associan is an AI robotic scientist built specifically for life scientists — enabling truly hands-free, 24/7 autonomous cell culture. At its core, the Associan is built on Teknician™, our proprietary AI-powered liquid handling platform designed for cell culture flasks and well plates.

The system integrates every component a cell culture lab needs:

• Automated incubator — CO₂-controlled environment with robotic access
• Refrigerator & heat bath — reagent storage and pre-warming
• 3-axis gantry robot — high-precision XYZ motion for flask transport
• Gripper system — multi-DOF flask handling and cap control
• Festo pump — precision liquid aspiration and dispensing
• IOLight microscope — real-time confluency imaging
• Waste management — fully closed loop, no manual intervention

Key specifications (as shown on the booth sign):

The Associan supports 45+ flask formats across multiple manufacturers and well plates, and can scale to:

• >1 billion cells per run from a single expansion campaign, OR
• Parallel expansion of 45+ distinct cell lines simultaneously

Scientists interact with the system through natural language — describing their protocol via speech or text — and the AI assistant translates that into precise robotic instructions. Real-time confluency imaging and cell count analysis ensure the system acts at the right biological moment, not on a fixed schedule.

Software Architecture

As Head of Robotic Engineer, I designed and implemented the entire software stack for the Associan. The system spans five architectural layers, from the desktop client to the physical hardware, with cloud integration for AI vision and data storage.

This work is a close collaboration with Enyuare, our hardware and integration partner, who built the physical hardware platform that my software stack runs on top of.

High-Level System Architecture

graph LR
    subgraph CLIENT["Client"]
        APP[Desktop App<br/>Electron + React]
        AI[AI Assistant<br/>LLM Service]
    end
    subgraph GATEWAY["Cloud Gateway"]
        SPRING[Spring Boot API<br/>Auth + Database]
    end
    subgraph ROBOT["Robot System"]
        NGINX[Nginx Proxy]
        SERVICES[Microservices<br/>Robot / Incubator<br/>Microscope / Fluid]
        ROS2[ROS2 + Jetson Orin]
    end
    subgraph HW["Hardware"]
        DEVICES[Gantry · Pump<br/>Flask Holders · Microscope]
    end
    subgraph CLOUD["Cloud"]
        GCS[Google Cloud<br/>Storage + Vision AI]
    end

    CLIENT -->|HTTP REST| GATEWAY
    GATEWAY -->|HTTP Forward| NGINX
    GATEWAY --> GCS
    NGINX --> SERVICES
    SERVICES --> ROS2
    ROS2 -->|CAN / Modbus TCP / Serial| HW
    SERVICES -.->|Images| GCS

    style CLIENT fill:#e1f5ff
    style GATEWAY fill:#fff4e1
    style ROBOT fill:#ffe1f5
    style HW fill:#f5e1ff
    style CLOUD fill:#e1ffe8

Layer breakdown:

• Desktop App (Electron + React): Cross-platform GUI for protocol setup, live monitoring, and result visualization. The AI Assistant layer accepts speech or text input and uses an LLM service to generate structured protocol definitions.
• Cloud Gateway (Spring Boot): Handles authentication, user management, database persistence, and forwards robot commands to the on-device proxy. Also coordinates image uploads to Google Cloud.
• Robot System (Nginx + Microservices + ROS2): On-device services running on the Jetson Orin. Nginx routes requests to domain-specific microservices (robot motion, incubator control, microscope, fluid handling). ROS2 nodes translate high-level commands to hardware control.
• Hardware Layer: Physical devices controlled via CAN bus, Modbus TCP, Dynamixel SDK, and HTTP/Serial.
• Cloud (Google Cloud): Stores microscope images and runs Vision AI for confluency analysis and cell counting.

Autonomous Cell Culture Workflow

The following flowchart illustrates what the Associan does end-to-end — from scientist intent to biological outcome:

flowchart TD
    START([Scientist Initiates Protocol<br/>via Desktop App / Voice Command]) --> LLM_PARSE[LLM Parses Instructions<br/>& Generates Protocol]
    LLM_PARSE --> RETRIEVE[Robot Retrieves Flask<br/>from Incubator]
    RETRIEVE --> IMAGE[Microscope Captures Image<br/>& AI Analyzes Confluency]
    IMAGE --> DECISION{Confluency ≥ Threshold?}
    DECISION -->|No| RETURN_INC[Return Flask to Incubator<br/>& Schedule Next Check]
    RETURN_INC --> RETRIEVE
    DECISION -->|Yes| ASPIRATE[Aspirate Old Media<br/>via Festo Pump]
    ASPIRATE --> WASH[Add PBS Wash Solution<br/>from Fluid Storage]
    WASH --> TRYPSIN[Add Trypsin<br/>Incubate at 37°C]
    TRYPSIN --> NEUTRALIZE[Add Neutralization Media]
    NEUTRALIZE --> SPLIT[Distribute to New Flasks<br/>via Flask Holder / Cap Control]
    SPLIT --> FRESH_MEDIA[Add Fresh Media<br/>to Each New Flask]
    FRESH_MEDIA --> STORE[Return Flasks to Incubator<br/>& Log Results to Cloud]
    STORE --> NOTIFY([Notify Scientist<br/>of Completion])

    style START fill:#e1f5ff
    style NOTIFY fill:#e1ffe8
    style DECISION fill:#fff4e1

Every step — from retrieving flasks, imaging cells, making passage decisions, to dispensing reagents and returning flasks — is executed autonomously. Scientists are notified on completion and can review results (cell counts, confluency trends, images) in the desktop app.

Hardware Device Communication

The robot-side communication layer maps each hardware device to its appropriate bus protocol:

graph TB
    subgraph COMP["Compute"]
        JETSON[Jetson Orin<br/>ROS2 Nodes]
    end
    subgraph BUS["Communication Buses"]
        CAN[CAN Bus]
        MODBUS[Modbus TCP]
        DYNAMIXEL[Dynamixel SDK]
        HTTP_HW[HTTP / Serial]
    end
    subgraph HW["Physical Hardware"]
        GANTRY[3-Axis Gantry<br/>XYZ Motion]
        PUMP[Festo Pump<br/>Liquid Handling]
        FLASK_H[Flask Holders<br/>6-DOF Grip]
        CAP_C[Cap Controller<br/>Open / Close]
        MICRO_KNOB[Microscope Focus Knob<br/>Auto-Focus]
        IOLIGHT[IOLight Microscope<br/>Imaging]
    end

    JETSON -->|CAN| CAN
    JETSON -->|Modbus TCP| MODBUS
    JETSON -->|Dynamixel| DYNAMIXEL
    JETSON -->|HTTP / Serial| HTTP_HW

    CAN --> GANTRY
    CAN --> FLASK_H
    CAN --> CAP_C
    MODBUS --> PUMP
    DYNAMIXEL --> MICRO_KNOB
    HTTP_HW --> IOLIGHT

    style COMP fill:#e1ffe1
    style BUS fill:#fff4e1
    style HW fill:#f5e1ff

The NVIDIA Jetson Orin serves as the on-device compute brain, running all ROS2 nodes. High-precision motion (gantry, flask holders, cap controller) runs over CAN bus for deterministic timing. The Festo pump uses Modbus TCP for precise volumetric control. Microscope focus is driven by Dynamixel servos, and the IOLight imaging unit communicates over HTTP.

The Engineering Journey: Build → Pack → Ship → Setup

Getting a precision robotic system from our lab to the SLAS 2026 expo floor in San Diego involved as much engineering as building it.

Packing and Shipping

The Associan is a large, sensitive system with a 3-axis gantry, precision liquid-handling components, glassware, and an automated incubator. Every joint, every pump line, and every optical component needed to survive a cross-country truck shipment without losing calibration.

We secured the system to a standard pallet with industrial ratchet straps, wrapped exposed components in moving blankets and foam, and documented the configuration to enable rapid reassembly on-site.

On-Site Setup

Arriving at the SLAS 2026 expo hall, we found our pallet in an empty hall before the crowds arrived. The unboxing process — unwrapping, positioning, reconnecting pneumatic and electrical lines, and recalibrating the gantry — took several hours of focused work.

After physical assembly, the software stack came up cleanly: the Jetson booted, ROS2 nodes initialized, and all microservices connected. We ran calibration routines for the gantry home position, flask holder alignment, and microscope focus. By the time the expo doors opened, the Associan was running live cell passaging demonstrations.

SLAS 2026 Expo

SLAS (Society for Laboratory Automation and Screening) is the premier conference for laboratory automation professionals, attracting thousands of researchers, engineers, and industry leaders from across the life sciences.

Over three days of live demos, the Whelix booth became a gathering point for anyone interested in the future of autonomous biology.

By the numbers:

• 3 days of continuous live demonstrations
• 1,000+ attendees engaged with the booth
• Dozens of in-depth technical conversations with researchers, engineers, and executives

Live demonstrations included autonomous cell passaging (flask retrieval → imaging → media exchange → splitting → return) and real-time confluency analysis — showing the full end-to-end loop with no human intervention.

Notable Visitors

The Associan attracted serious interest from organizations across the life science and industrial automation space:

Festo — A global leader in industrial and process automation. Their interest centered on the robotic integration layer and potential for their pneumatic and motion components to be incorporated into future Associan configurations.

ThermoFisher Scientific — The world’s largest supplier of scientific instrumentation and life science reagents. The conversation explored how the Associan could accelerate workflows that today require ThermoFisher benchtop equipment operated manually.

Boston Children’s Hospital — Pediatric research teams are among the most resource-constrained in academic biology. Their interest focused on using the Associan to enable small teams to run large-scale cell culture experiments that would otherwise require dedicated technician time around the clock.

Additional visitors included biotech companies, contract research organizations (CROs), academic research groups, and automation engineers — representing nearly every sector that touches cell culture today.

Impact and Significance

Cell culture is the backbone of biological research and manufacturing — from developing cancer therapies and regenerative medicine to producing recombinant proteins and viral vectors. Yet it remains one of the most labor-intensive, error-prone, and time-constrained tasks in any life science lab.

The Associan changes this equation:

• A scientist who previously spent 2–4 hours daily on cell maintenance can now redirect that time entirely to experimental design and analysis
• A single operator can oversee 4–5 Associan robots simultaneously, dramatically multiplying research throughput
• The >99% execution consistency metric means variability from human technique is effectively eliminated
• 24/7 operation means cell passaging happens at the biologically optimal moment — not when a technician is available

One researcher from CMU noted that the system could allow a single graduate student to manage cell culture workloads that previously required a dedicated lab technician.

What’s Next

The SLAS 2026 debut marks the beginning of Associan’s presence in the life science automation market. The conversations and partnerships initiated on the expo floor are continuing into formal pilot programs and integration discussions.

If you’re interested in learning more about the Whelix Associan or exploring how it might fit your lab’s needs, visit whelix.co/products#associan.