QPillars

QPillars builds AI agents for controlling laboratory instruments through natural language, with human approval built into every step. Founded in 2024 in Zurich, the company pairs competitive programming expertise—the team includes four National Olympiad medalists in Informatics and Physics—with hands-on experience shipping instrument control software for diagnostic platforms. They embed with clients to solve specific problems: MCP servers that let AI communicate with lab hardware, control systems that run on real instruments in regulated environments, and LiquidBridge, a digital twin platform for liquid handling robots. The engineering office sits in Chisinau, Moldova. Six engineers, no management layers between problem and solution.

Services and capabilities

QPillars works across three main offerings. Agentic AI for Instruments builds MCP servers and natural language interfaces using CrewAI and FastAPI. Scientists describe intent; the AI plans steps; humans review and approve; the instrument executes. A full audit trail follows. The company has shipped 5+ instrument platforms and extracted 47 instrument primitives from single-sentence descriptions.

Systems Engineering covers the full stack: firmware, real-time data pipelines, embedded control systems in C++ and Rust, and regulatory-grade code. They handle IVD compliance and ISO 13485 regulated software. The team has deep experience with diagnostic platforms where real-time constraints matter—coordinating 96-channel pipetting, for instance.

LiquidBridge Digital Twin simulates liquid handling protocols before execution, optimizes with AI agents, and tracks results. It supports Hamilton STAR, Tecan FluentControl, Beckman Biomek, and others without vendor lock-in. The platform catches errors like tip collisions and insufficient volumes before reagents are wasted. Industry data cited on their site shows uncalibrated protocols cause 20–30% reagent over-dispensing; LiquidBridge is in active development with a beta launch targeted for Q2 2026. Early access is free; five lab slots remain.

Notable work

Lino Biotech AG engaged QPillars as an embedded development team to strengthen instrument control software for advanced detection instruments. Dr. Andreas Frutiger, VP of R&D, noted their depth of understanding transformed product development.

Ridgeview Instruments AB partnered with QPillars to build a cloud-connected remote monitoring platform for LigandTracer instruments. The platform enables real-time data visualization and intelligent alerting so researchers can track experiments remotely. CEO Jos Buijs highlighted their rapid problem understanding and agile execution.

LiquidBridge (internal R&D) demonstrates their core technical approach: connecting AI agents to lab instruments safely through MCP, CrewAI, and a planner-executor architecture. Iacob Marian, Technical Lead and co-founder, calls it proof that "agentic AI can control lab instruments safely through an open standard."

How they work

Engagements start with a 2-week Discovery Sprint. QPillars embeds with the client team, audits the current instrument software landscape, and delivers a technical roadmap. Full integrations typically span 2 to 6 months depending on complexity and regulatory requirements. They work with biotech, medtech, diagnostics (IVD), and pharma. Integration occurs on top of existing LIMS, ELN, and vendor software—not replacing it. The engineering team writes the code; no sales layer stands between client and builder.

Team and credentials

Iacob (Technical Lead, Informatics Olympiad medalist, 10 years in software engineering, 4 years in IVD instrument control) and Tatiana (Business Development, Physics Olympiad medalist, ETH Master in Integrated Building Systems) founded the company. The six-person team includes Stefan, Mihai, and Sevastian—all Olympiad medalists—plus Adelina (UI/UX, designing interfaces for complex lab workflows) and Vasile (Operations). The advisory board includes Dr. Sc. Andreea Loredana Cretu, Head of BioMed Entrepreneurship Programs at the University of Zurich. Backed by Innosuisse (Swiss Innovation Agency), Kommerskollegium/Techstep (Swedish National Board of Trade), and Invest Moldova. They work in Python (FastAPI), TypeScript/React, Rust, and C++. MCP protocol and CrewAI for agent orchestration.