--- title: "Research, platform engineering, and aerial robotics | Peinser" description: "Peinser helps teams resolve difficult research, platform, and UAV engineering problems, then turn the right answer into a system they can operate." canonical: "https://peinser.com/" --- EST. 2021 · Antwerp, Brussels, Leuven, Liège & Remote · Now discussing selected Q4 2026 and Q1 2027 engagements # Turn hard engineering questions into working systems. Peinser is a European engineering team for work that crosses research, platform engineering, and aerial robotics. We help you choose the right approach, prove it, and put it into operation. [Discuss your project](https://peinser.com/#contact) [Explore Flightdeck →](https://peinser.com/flightdeck) Flightdeck · Design-partner alpha · Browser-native UAV operations 01 — Practice ## Research & development. Applied research, machine learning, and engineering R&D for questions without an off-the-shelf answer. We produce reproducible prototypes, peer-reviewable results, and a clear path from evidence to production. Applied ML · Data & modelling · Lit reviews · Co-publications 02 — Practice ## Platform engineering. Distributed systems, cloud infrastructure, CI/CD, and observability for teams whose platform has become a delivery constraint. We design the target, ship the first production slice, and leave your team able to run it. Distributed Systems · IaC · Kubernetes · CI/CD · SRE & Observability 03 — Practice ## UAV operations software. Pilot workstations, mission planning, field tooling, payload integration, and autonomy engineering for custom fixed-wing, VTOL, and multirotor systems. Flightdeck · MAVLink · ArduPilot · PX4 integration · RTK · Video · Perception 04 — Practice ## Strategy & direction. Technical audits, discovery, and decision memos for teams facing an expensive or difficult-to-reverse choice. We make the options, evidence, costs, and risks explicit before you commit. Technical audits · Decision memos · Roadmaps · Second opinions — 01 / Robotics ## A pilot workstation for custom UAVs. Flightdeck brings terrain-aware mission planning, live flight operations, video, traffic, RTK, radio, and vehicle tooling into one private browser workspace. ![Flightdeck planning a terrain-relative fixed-wing survey, with the estimated flown path and terrain altitude profile visible.](https://peinser.com/flightdeck-planner.webp) Flightdeck terrain-relative survey planning, estimated flight path, and altitude profile. — Flightdeck / Design-partner alpha ### Built around the aircraft and the operator. Generic ground stations are broad by design. We configure Flightdeck around the airframe, autopilot, payload, communications, and operating constraints your team actually uses. - **Plan with aircraft context**: Terrain-relative surveys, QGroundControl interchange, geofences, rally points, and estimated flown paths rather than only straight waypoint lines. - **Operate from one cockpit**: Primary flight instruments, mission progress, tactical map actions, manual control, video, traffic, weather, and acknowledged command workflows. - **Work in the field**: Firmware flashing, SiK radio configuration, u-blox RTK survey-in, parameter management, MAVLink inspection, and vehicle-file access in the browser. - **Control who may command**: Production operators authenticate with private-PKI identities. An exclusive identity-scoped write lease gates operator traffic sent toward the aircraft. - **Integrate perception and autonomy**: Browser and edge inference, custom payloads, companion computers, and visual-guidance integrations around the camera and autopilot stack. - **Deploy privately**: One isolated Flightdeck runtime per vehicle, a QUIC rover link, and a self-hosted operator environment that can later be composed into Operator-managed operations. [Explore Flightdeck →](https://peinser.com/flightdeck) [Try the standalone planner →](https://flightdeck.uas.peinser.com) — 02 / Method ## Start with a focused six weeks. Long enough to resolve the important uncertainty. Short enough to act on the result. We agree the question and decision criteria up front, then finish with evidence, a decision, and a working first slice where appropriate. 01 — Phase ### Discovery. Read the literature, the code, the data, and the runbooks. Talk to operators and end users. The deliverable is a findings memo with explicit assumptions and open questions. WK 1–2 · Findings memo 02 — Phase ### Decision. A small set of options, each with a cost, a risk profile, and an owner. We choose together; the choice — and the rationale — is the deliverable. WK 3–4 · Decision memo 03 — Phase ### Ship. Put the decision in front of real users, real infrastructure, or a representative flight environment. The output is a thin, instrumented slice and a concrete plan for what follows. WK 5–6 · Working first slice > The hardest engineering decisions are rarely about what to build. They are about what to keep running, what to deprecate, and which version is worth the cost of operating. — Internal note · Wk 03 · Aerial inspections engagement — 03 / Projects ## Products and platforms. Products we build from recurring engineering problems. Try the public tools, inspect the open-source work, or talk to us about deploying them in your operation. Open source · Geodesy Live · v0.8 ### CORSHub A federated CORS / RTK network for centimetre-grade GNSS. An auditable community network of base stations that broadcasts RTCM corrections to anything with an RTK receiver; surveyors, agricultural rovers, and our own drones. Open protocol, open data, federated by design. github.com/peinser/corshub · MIT · NTRIPv2 · Open Policy Agent [Project site →](https://github.com/peinser/corshub) Open source · Networking Pre-1.0 · MVP ### quic-sni-router A small QUIC SNI router for pod-terminated mTLS and HTTP/3 services. Reads the SNI from incoming QUIC v1 and v2 packets and forwards the original datagrams, untouched, to the right backend pod. TLS, mTLS, and HTTP/3 termination stay in the application; the router is one UDP ingress in front of many QUIC services. Built for Kubernetes, configured in YAML, fuzzed and containerised. github.com/peinser/quic-sni-router · Apache 2.0 · QUIC v1/v2 · HTTP/3 [Project site →](https://github.com/peinser/quic-sni-router) Product · Documents Live · v1.0 ### Xarta A document protocol for the full lifecycle: render, sign, archive, dispatch. Teams repeatedly rebuild the same document plumbing: rendering PDFs from templates, versioning templates, signing output, assembling bundles, dispatching Peppol invoices, and archiving records for the required retention period. Xarta consolidates the lot behind a DAG protocol. The submitter declares the flow; Xarta executes it deterministically. Plug-in template engines, templates versioned in git rather than locked in a vendor UI, Peppol BIS 3.0 in the same pipeline as everything else. On-prem by default. On-prem · Peppol BIS 3.0 · Document Lifecycle · Template Rendering · PDF/A · PAdES Spin-out · Energy · EV Charging In market ### Zonnetanken Smart EV charging that follows the sun. An open charging network with auditable sessions for employee cost recovery. Started as a side project; now an independent product with paying users across Belgium. Its OCPP interface gives operators control of their charging infrastructure, data, and workflows. zonnetanken.be · Web · OCPP · Alfen · Peblar [Visit Zonnetanken →](https://zonnetanken.be) — Selected clients Acerta · ULiège · Innovatix · Duvel Logistics · H.Essers · Van Cleven NV — 04 / Research ## Research and open source. Selected publications and code we have shared with the community. Most of this work feeds back into client engagements. Research paper · ICML 2020 ### Likelihood-free MCMC with amortized approximate ratio estimators. - Machine learning - Simulation-based inference - MCMC proceedings.mlr.press ↗ Research paper · NeurIPS 2022 ### Towards reliable simulation-based inference with balanced neural ratio estimation. - Machine learning - Bayesian inference - NeurIPS papers.nips.cc ↗ Research paper · TMLR 2022 ### A crisis in simulation-based inference? Beware, your posterior approximations can be unfaithful. - Machine learning - Simulation-based inference - Calibration openreview.net ↗ Research paper · MNRAS 2022 ### Constraining warm dark matter with stellar streams and neural networks. - Machine learning - Bayesian inference - Astronomy academic.oup.com ↗ Code repository · amortized-experimental-design ### Learning to select optimal experiments. - Statistics - Machine learning - Tutorial - PyTorch github.com/JoeriHermans ↗ Code repository · esp-remoteid ### ESP32-S3 OpenDroneID broadcast firmware for EU Direct Remote ID integrations. - Remote ID - EU 2019/945 - OpenDroneID - Ed25519 github.com/peinser ↗ Now discussing selected Q4 2026 and Q1 2027 engagements ## Bring us the problem before the specification. Tell us what is uncertain, what you have tried, and what a useful outcome would change. In the first conversation, we will test whether we are the right team and outline how we would approach the work. No prepared brief required. [Email your project context](mailto:hello@peinser.com?subject=Project%20conversation)