I'm Chris von Csefalvay, a Distinguished Engineer at HCLTech's Robotics Intelligence CoE, specialising in sensing and reasoning for embodied AI. My work brings together machine perception and world models, focusing on post-training to develop the reasoning that connects sensing to action, particularly in robotic surgery and other high-stakes domains.

With Tamás Földi, I am a principal investigator of the Asset Factory Blueprint, an open-source reference implementation from HCLTech's Robotics Intelligence CoE for turning photographs, scans and CAD into reproducible OpenUSD assets and reinforcement learning environments. I am also one of the creators of CERES, an open-source application that turns every Meta Quest headset into a high-quality capture device for egocentric hand-object interaction data.

My book The Craft of Post-Training: A Practical Guide for AI Engineers and Developers was published by No Starch Press in 2026 — a practical guide to turning foundation models into production-ready systems, from SFT and RLVR through GRPO, DPO, LoRA and agentic fine-tuning. I am also the author of Differentiable Photon Transport, Volume I: Foundations, an illustrated, executable introduction to differentiable X-ray image formation and inverse problems. You might also know me from the world's largest collection of Covid-19 data, the public data set on the 2019 Samoan measles outbreak, the public data set on the 2018 West African EBOV outbreak, the first community-driven book on Julia, and my monograph Computational Modeling of Infectious Disease (Elsevier, 2023). I write about things here — lately, mostly about post-training, agentic systems, and the gap between AI hype and AI craft.

I'm a computational epidemiologist and data scientist with a primary focus on AI/ML: post-training methods, agentic system design, and the practical problem of making foundation models reliable in high-stakes domains. I am the author of The Craft of Post-Training, Differentiable Photon Transport, Volume I: Foundations and Computational Modeling of Infectious Disease, as well as a growing body of research.

I attended the University of Oxford for my undergraduate studies, graduating with a distinction (1st class), and returned there for my graduate studies. I also hold degrees from Cardiff and Robert Gordon University Aberdeen, and have studied at the University of Leiden under an Erasmus programme. Today, I am board certified in public health, a Fellow of the Royal Society for Public Health, a Senior Member of IEEE and a member of several other learned societies.

I am currently a Distinguished Engineer at HCLTech's Robotics Intelligence CoE, where I work on post-training for embodied sensing and reasoning. My primary focus is on the way world models and machine perception can create more reliable systems of physical AI. My current key research interest is the use of agentic intelligence in an embodied context (Agentic Robotics or Agentic Embodied Intelligence). I am also a visiting thesis supervisor for students on the data science track of the mathematics programme at the Budapest University of Technology and Economics.

As part of my service to the profession, I've been a reviewer for JOSS, Frontiers in Immunology and Frontiers in Medicine. Between 2022 and 2023, I served Virginia as a member of the Medical Reserve Corps, and currently serve in both the Rocky Mountain MRC of Denver and Colorado's radiation response MRC. I was also part of the COVID-19 Healthcare Coalition and the Covid Tracking Project.

You can find a list of my publications here. My work on COVID-19 has been featured in the media. I also write on various topics as the fancy strikes — you can peruse my notebook here.

• AI/ML: post-training methods (SFT, RLVR, GRPO, DPO, ORPO), agentic AI ecosystems, agentic fine-tuning, preference optimisation, emergence, simulation agents and hybrid simulact networks
• Computational epidemiology: agent-based modelling, disease-avoidant behaviours, epidemics over dynamic networks, geospatial computational epidemiology — all primarily in the context of infectious diseases
• Pharmacovigilance: VAERS, passive reporting, text mining of large complex report sets using LLMs
• Computational dynamics: computational dynamics of complex systems, computational dynamics of infectious diseases, computational dynamics of social systems
• Public health: evidence-based public health policy, public health ethics, public health law, quality and systems improvement in public health

I am a visiting thesis supervisor for students on the data science track of the mathematics programme at the Budapest University of Technology and Economics — you can read more about this here.

From time to time, I get invited to give talks and lectures. My capacity for this is somewhat limited due to other commitments, but if you are interested in having me speak at your event, please get in touch.

Tocilizumab in addition to standard of care in the management of COVID-19: A meta-analysis of RCTs. Mutua V, Henry BM, von Csefalvay, C, Cheruiyot I, Vikse J, Lippi G, Bundi B, Mong'are N. Acta Biomed. 2022, 93(1):e2022014. doi: 10.23750/abm.v93i1.12208.

JAMPI: Efficient matrix multiplication in Spark using barrier execution mode. Foldi T, von Csefalvay C, Perez NA. BDCC 2020, 4(4):32. doi: 10.3390/bdcc4040032.

VAERS data reveals no increased risk of neuroautoimmune adverse events from COVID- vaccines. von Csefalvay, C. medRxiv 2021. doi: 10.1101/2021.06.13.21258851.

Tuberculous polyserositis in endemic areas with an emphasis on empiric therapy: A case report. Munguti J, Mutua V, Cheruiyot I, von Csefalvay C, Opare-Addo P, Kiko N, Wanjiru R. Medicine: Case Reports. 2022, 3(4):e0221. doi: 10.1097/MD9.0000000000000221.

See more here.

The OR Society | Member since 2020

Royal Society for Public Health | Fellow since 2021

TOPRA | Registered Member since 2021

American Public Health Association | Member since 2021

I live in Denver, CO, with my Golden Retriever, Oliver. Born in Budapest, Hungary, I've spent my formative years in England, but have called the Netherlands, Germany, Belgium, Austria and the United States home at various points in my life. Since 2020, I've been living in the United States, where I am a permanent resident.

In my free time, I enjoy reading (mostly Cold War history and classical Greek drama), cooking (with not a lot of success) and Renaissance polyphony (esp. Thomas Tallis and Orlando di Lasso). I'm credited for a somewhat pretty sequence of integers called Jellyfish Heart numbers (OEIS A344856), which I discovered in 2020 while doing something that has nothing at all to do with my day job.

I've lived with neuromyelitis optica since 2015, which landed me in a wheelchair but never slowed me down. It did, however, give my work a personal stake: to find ways in which machine intelligence and sensing can expand what people can do, see and perceive, on their own terms. My experience of spinal cord injury has shaped my interest in AI not as an academic pursuit for its own sake but as a means of extending human agency, independence and capability.

In what little is left of my free time, I'm an adaptive multisport athlete, competing in wheelchair rugby, adaptive rowing (category PR1) and the SkiErg (category SIT2). In the latter, I currently hold the world records for 100m, 500m, 2,000m, 6,000m, 21,097m and 42,195m distances and the 60 minutes time. I occasionally rant and whine about my forays in the world of adaptive exercise and my attempts at doing so in a mostly science-based manner.

My last name rhymes with Chick-Fil-A.