Intelligent Software for Scientific Instruments

Orchestration platforms (lab OS, workflow engines) only work when each instrument exposes a usable interface. We build that interface - on the instrument, in the languages the firmware already runs, against the standards platforms consume (SiLA 2 and MCP). You can plug our work into any orchestrator, or skip the orchestrator entirely and let agents talk to the instrument directly.

Productized engagements run 4-12 weeks against fixed scope and fixed price - MCP server builds, AI-readiness assessments, specific instrument-software deliveries. Embedded engagements (legacy modernization, v2 greenfield, multi-instrument programs) start at 3 months and extend month by month. We always start with a 30-minute technical call with the founder to scope honestly before any contract.

Architecturally yes - audit trails, role-based access, time-correlated logging, immutable history, validation impact assessments. One of our co-founders has shipped C++ instrument-control software on a clinical-grade IVD platform. Validation evidence is the customer's QMS work; we ship the architecture, documentation, and test artifacts the QMS expects. QMS audits and regulatory consulting sit outside our scope - we partner with QMS specialists when those are part of a customer's roadmap.

Python (FastAPI, AI agents, instrument SDKs), TypeScript / React / Next.js (operator UI, dashboards), and Rust/C++ (host control where the instrument needs it). For new IEC 62304 Class C programs, we have shipped Rust greenfield communication libraries (Ferrocene's Class C qualification, Jan 2025, makes Rust legally usable in regulated medical software). Our instrument layer supports any wire protocol - serial, USB, CAN, GPIB, REST, gRPC, SiLA 2.

QPillars is a Swiss GmbH headquartered in Zurich, with a senior engineering hub in Chișinău, Moldova (MITP). One Slack channel, one delivery team, one accountable founder on every engagement.