Konstantin D. Pandl

About

I am a protocol engineer at Provable, working on scalable blockchain systems. Furthermore, I did work on machine learning model inference in zero knowledge proofs. Before that, I did my Ph.D. at KIT in Germany, advised by Ali Sunyaev. Over the summer of 2022, I was a visiting researcher at the Systopia lab of UBC Vancouver, hosted by Ivan Beschastnikh. Previously, I graduated in Electrical and Computer Engineering.

My research interests are in privacy-preserving machine learning, and blockchain-based data marketplaces for AI.

Crypto Work

Selected engineering work around zero-knowledge and the Aleo blockchain:

• zkML transpiler: a Python→Leo transpiler that turns ML models into Leo code, lets you execute the transpiled code from Python, and generates zk proofs.
• snarkOS scaling: performance and reliability work across proving and networking paths aimed at lower latency and higher throughput.
• Know Your Address (KYA): a ZK pattern on Aleo that proves a human-not just a private key or unsupervised agent-approved an action, with client-side biometrics and on-chain proof verification (demo).

Publications

• Scalable Data Point Valuation in Decentralized Learning (arXiv preprint)
  Konstantin D. Pandl, Chun-Yin Huang, Ivan Beschastnikh, Xiaoxiao Li, Scott Thiebes, Ali Sunyaev
• Reward Systems for Trustworthy Medical Federated Learning (ACM Transactions on Computing for Healthcare, June 2025)
  Konstantin D. Pandl, Florian Leiser, Scott Thiebes, Ali Sunyaev
• Artificial Intelligence as a Service (Business & Information Systems Engineering)
  Sebastian Lins, Konstantin D. Pandl, Heiner Teigeler, Scott Thiebes, Calvin Bayer, Ali Sunyaev
• How detection ranges and usage stops impact digital contact tracing effectiveness for COVID-19 (Scientific Reports)
  Konstantin D. Pandl, Scott Thiebes, Manuel Schmidt-Kraepelin, Ali Sunyaev
• Trustworthy machine learning for health care: scalable data valuation with the shapley value (Proceedings of the ACM Conference on Health, Inference, and Learning (CHIL 2021), Virtual Event, Apr 8 - 9)
  Konstantin D. Pandl, Fabian Feiland, Scott Thiebes, Ali Sunyaev
• Drivers and Inhibitors for Organizations’ Intention to Adopt Artificial Intelligence as a Service (Proceedings of the 54th Hawaii International Conference on System Sciences (HICSS 2021), Hawaii, Jan 5 - 8)
  Konstantin D. Pandl, Heiner Teigeler, Sebastian Lins, Scott Thiebes, Ali Sunyaev
• What Your Radiologist Might Be Missing: Using Machine Learning to Identify Mislabeled Instances of X-ray Images (Proceedings of the 54th Hawaii International Conference on System Sciences (HICSS 2021), Hawaii, Jan 5 - 8)
  Tim Rädsch, Sven Eckhardt, Florian Leiser, Konstantin D. Pandl, Scott Thiebes, Ali Sunyaev
• On the Convergence of Artificial Intelligence and Distributed Ledger Technology: A Scoping Review and Future Research Agenda (IEEE Access 8, 57075-57095, March 2020)
  Konstantin D. Pandl, Scott Thiebes, Manuel Schmidt-Kraepelin, Ali Sunyaev
• A Cost and Time Efficient Approach for Storing and Querying Genomic Data in Ethereum Smart Contracts (iDash Privacy & Security Workshop, October 2019)
  by Mikael Beyene, Niclas Kannengießer, Konstantin D. Pandl, Scott Thiebes, Ali Sunyaev
• Optimization of System Parameters for an Online Driving Style Recognition (IEEE International Conference on Intelligent Transportation Systems, November 2016)
  by Dominik Dörr, Konstantin D. Pandl, Frank Gauterin

Talks

• cii lab@KIT, Federated Learning and Data Valuation, Current Research (University of British Columbia, Vancouver, Canada, May 2022)
  
• Convergence of AI and DLT – review and future research agenda [Slides] | BloG3 research project - blockchain-based health data management (Oasis Labs, San Francisco, CA, USA, April 2020)
  
• How the Convergence of Blockchain and Artificial Intelligence will Benefit Healthcare Researchers (Blockchain for Science Conference, Berlin, November 2019)
  [Video] [Abstract]

Indie-Hacking

Besides crypto work and research, I like to stay current with large ML models by building small products end-to-end. During my time in academia, while writing our HICSS paper on AI as a Service (AIaaS) adoption, I had to transcribe and analyze long interviews and learned first-hand how critical good transcripts are. That experience nudged me to put AIaaS into practice and ship my own service: InstantTranscriber - see the academic transcription workflow - a pragmatic speech-to-text tool focused on fast, accurate transcripts for interviews and podcasts.