CV
Sidharth Kannan
Machine learning researcher focusing on agents for science and engineering
Summary
Research Engineer at ChipAgents building efficient agents for electronic design automation.
Education
- B.S. in Computer Science and B.S. in PhysicsMarch 2026University of California, Santa BarbaraGPA: 3.94, High Honors
Work Experience
- Research EngineerMarch 2026 - PresentChipAgentsBuilding efficient machine learning agents for electronic design automation.
- Built domain-specific datasets for semiconductor tapeout and post-training of GLM 5.2.
- Shipped an agentic analog mixed-signal design verification toolchain used by tier-one semiconductor companies.
- Reduced token use by 40% through model routing, programmatic tool calling, and caching optimizations.
- Computational Physics FellowJune 2025 - March 2026Los Alamos National LaboratoryDeveloped computational models for relativistic electron scattering and fusion-reactor simulations.
- Developed codes for modeling relativistic electron scattering in dilute gases.
- Built flow-based models to super-resolve observables in particle-in-cell simulations.
- Undergraduate Research AssistantOctober 2022 - PresentUniversity of California, Santa BarbaraConducted research in generative machine learning, scientific computing, and probabilistic hardware.
- Built flow-based generative models for representation learning of cosmological fields.
- Built the first hardware implementation of a higher-order Ising machine.
- Increased device capacity by 20% through graph-clustering algorithms for resource sharing.
- Research InternJune 2024 - December 2024LIGO LaboratoryDeveloped graph-neural-network surrogate models for interferometer simulation.
- Achieved up to 800x speedups over traditional numerical methods with minimal loss of fidelity.
- Improved thermal-aberration simulation by comparing frequency- and spatial-domain methods.
- Flight Software InternJune 2023 - September 2023Astranis Space TechnologiesBuilt satellite telemetry analysis tools and a GNSS receiver interface.
- Firmware Engineering InternJune 2022 - September 2022TenaFe, Inc.Developed a software framework and simulation modules for SSD controller verification.
Skills
Machine Learning
- PyTorch
- JAX
- Slime
- Harness Engineering
- Flow Matching
- Weights & Biases
Scientific Computing
- NumPy
- SciPy
- Distributed Training (DDP)
- SLURM
Chip Design
- FPGA
- Vivado
- SystemVerilog
- SPICE
- Agents for Chip Design
Publications
- Lambda-CFM: Scale-Aware Representation Learning for Cosmology with Flow Matching2025
- Graph Neural Networks for Interferometer Simulation2025NeurIPS 2025 Workshop on AI for Science
- All-to-all Reconfigurability with Sparse and Higher-Order Ising Machines2024
- The Case for Data Centre Hyperloops202451st ACM/IEEE International Symposium on Computer Architecture
Presentations
- Lambda-CFM: Scale-Aware Representation Learning for Cosmology with Conditional Flow Matching2025Simons Foundation, Learning the Universe CollaborationInvited talk
- Self-Consistent Relativistic Electron Scattering for X-Ray Diagnostics2025Kavli Institute for Theoretical Physics, Undergraduate Research SymposiumSanta Barbara, CA
- Statistical Mechanics of Machine Learning2025NSF Institute for Artificial Intelligence and Fundamental Interactions Reading Group
- Graph Neural Networks for Interferometer Simulations2024Kavli Institute for Theoretical Physics, Undergraduate Research SymposiumSanta Barbara, CA
Teaching
- Undergraduate Learning Assistant2023University of California, Santa BarbaraRole: Teaching assistantIntro to Scientific Computing, Electrostatics, Data Structures and Algorithms I & II, and Discrete Mathematics.
- CMPTG CS 5: Statistical Physics and Neural Networks2025University of California, Santa BarbaraRole: Course instructorTen-week course connecting neural-network theory and statistical physics.
- PHYS CS 5: Non-equilibrium Statistical Mechanics2026University of California, Santa BarbaraRole: Course instructorTen-week course on stochastic processes, Brownian motion, kinetic theory, and open quantum systems.
Portfolio
- Mini-GPT2026
- Fermiacc2026Ai for scienceAI agents for autonomous theory generation and validation in high-energy physics.