CERN detector
Rajat Gupta

Rajat Gupta

Research Fellow • University of Birmingham / CERN

Research Fellow at the University of Birmingham and CERN, working on machine-learning methods for real-time data processing in high-energy physics. My research focuses on anomaly detection, model distillation and data compression for systems with strict latency, memory and compute limits. Recent work includes a distilled variational-autoencoder method for real-time anomaly detection and an industry collaboration with HPE Labs on energy-efficient data compression.

Research Interests

Physics analysis, trigger/DAQ, and ML for real-time event selection.

BSM (Dark Matter / Invisible Higgs) Run-3 Triggers & Commissioning HL-LHC Trigger Upgrades Anomaly Detection (Distilled ML) Calorimeter Data Compression MC Generator Validation/Tuning Muon Detectors (GEM/RPC)

Key Contributions

Selected leadership and technical contributions across ATLAS and CMS.

Run-3 MET Trigger Commissioning (ATLAS)

Leading commissioning and physics-performance validation of calorimeter-based missing transverse momentum triggers under high pile-up conditions.

HL-LHC Global Trigger: Pile-up Suppression

Developing and validating pile-up suppression algorithms for next-generation trigger architectures, optimised for low latency and hardware constraints.

NomAD: Distilled Level-1 Anomaly Trigger

Developed a novel distilled ML anomaly-detection trigger for the ATLAS Level-1 Topological Trigger, enabling deployable real-time selection for rare signatures.

Industry Collaboration (HPE): Calorimeter Compression

Leading an industry–academia collaboration on ML-driven calorimeter data compression for future collider detectors, bridging physics constraints and hardware-aware ML design.

Selected Publications

Full list on INSPIRE.

Contact

For academic positions, collaborations, and invited talks.

Email: rajat.gupta@cern.ch
Location: Birmingham, United Kingdom / CERN, Geneva
CV: Download PDF
Last updated: September 2026