This site provides a llms.txt file at /llms.txt for AI chatbots and parsers.

I build AI systems from the environment up.

I'm Sebastian Boehler, a research engineer and founder working across high-speed simulation, trading infrastructure, autonomous systems, and source-grounded learning.

IEEE-published research · Founder at Sunderlabs · Co-founder and CTO at HB Capital · M.Sc. Computer Science at Tübingen

Selected work

Four systems, each presented with the evidence it has earned and the boundary it has not crossed.

Sunderlabs / FlightRL

Training compact policies before they reach real hardware.

High-throughput native environments, privileged teachers, edge-shaped policies, telemetry, and staged gates.

Evidence
FlightRL repository and Sunderlabs research program.
Boundary
Learned navigation is pre-deployment research, not live-control authority.

HB Capital

Keeping market claims attached to their evidence.

Market structure, positioning, macro context, native research tooling, and read-only account context.

Evidence
HB Capital's live research workspace.
Boundary
Research access grants no trading authority.

Dialogue research

Testing whether smaller language models can predict what users say next.

QLoRA next-turn prediction with multi-step dialogue rollout evaluation.

Evidence
Peer-reviewed IEEE proceedings paper.
Boundary
The publication supports its reported experiments, not broader model-performance claims.

LecturePilot

Turning private course material into controlled learning workspaces.

Professor-owned sources, enforced unlocks, typed tutor capabilities, and learner-owned work.

Evidence
LecturePilot's working source-grounded course system.
Boundary
Live-pilot status is not production-security or learning-efficacy approval.

Recent writing

Notes on research systems, engineering practice, and how models behave in the environments around them.

Browse all essays

About

I moved from self-taught builder and backend engineer to research engineer and founder.

Today I work through Sunderlabs and HB Capital while studying computer science at Tübingen.

Methods

  • High-throughput simulation and staged deployment gates
  • Research infrastructure with explicit evidence boundaries
  • Source-grounded learning systems and typed tutor capabilities
View CV

Timeline

  1. Founder & AI Engineer

    Sunderlabs

  2. Co-founder & CTO

    HB Capital

  3. M.Sc. Computer Science

    University of Tübingen

  4. B.Sc. Computer Science

    IU International University

  5. Backend Developer

    LI.FI

  6. Founder & Developer

    Boehler IT Solutions

  7. Backend Developer

    remotly

Contribution arc

A quiet record of sustained public building from 2017 onward.

  1. 2026

    6,002 contributions

  2. 2025

    1,798 contributions

  3. 2024

    1,145 contributions

  4. 2023

    256 contributions

  5. 2022

    524 contributions

  6. 2021

    554 contributions

  7. 2020

    574 contributions

  8. 2019

    313 contributions

  9. 2018

    1 contribution

  10. 2017

    2 contributions

Build the next system on evidence.

If the work above connects to a problem you are working on, let's discuss a collaboration.

Discuss a collaboration