Rishabh Verma

Embedded AI & Machine Learning Engineer

Freiburg, Germany

GitHubLinkedIn

Download CV (PDF)

Work Experience

2025 – present

Wissenschaftlicher Mitarbeiter (Research Associate)

Institute of Physiology, University of Freiburg

Freiburg, Germany

  • Run and maintain the two-photon imaging and Neuropixels analysis pipeline: Suite2p and Cellpose for extraction and segmentation, plus deep learning methods for 3D volume registration and microglia segmentation and tracking.
  • Design and build closed-loop behavioural rigs and two-photon imaging hardware: Blender VR maps, synchronised multi-camera Basler acquisition, microcontroller-based reward and stimulus control, Optotune tunable lens integration.

Oct 2022 – 2025

Wissenschaftliche Hilfskraft (Research Assistant)

Institute of Physiology, University of Freiburg

Freiburg, Germany

  • Developed acquisition and simulation tools for imaging and electrophysiology experiments, including hardware and software for two-photon and Neuropixels experiments.
  • Built VR systems for mice imaging and electrophysiological recording of neuronal and behavioural data.

2019 – 2022

Junior Research Fellow

Centre For Airborne Systems, DRDO

Bengaluru, India

  • Developed a single-shot CNN RADAR detector outperforming conventional techniques such as CFAR, and LSTM-based modular tracking robust to unknown clutter density and detection probability.
  • Published three peer-reviewed papers on radar detection, tracking and IFF mode code detection.

Education

2022 – 2025

M.Sc. Embedded Systems Engineering, specialisation in AI

Albert-Ludwigs-Universität Freiburg

Freiburg, Germany

  • Thesis: 3D Scene Graph Filtering using Graph Neural Networks.

2014 – 2018

B.Tech. Electronics and Communication Engineering

Jaypee Institute Of Information Technology

Noida, India

Internships

2017 – 2018

Research Intern

IIIT-Delhi

New Delhi, India

  • Implemented optimum frequency hopping strategy for wireless nodes giving better throughput while avoiding jammers as compared to the naive approach using reinforcement learning.

Skills

Machine Learning & Imaging

  • PyTorch
  • TensorFlow
  • Suite2p
  • Cellpose
  • 3D registration & segmentation

Programming & Tooling

  • Python
  • C/C++
  • MATLAB
  • Linux
  • Git
  • ROS
  • ARM Cortex-M
  • ESP32
  • Raspberry Pi

Projects

Learned 3D Registration for Volumetric Microscopy

Aligning two-photon and confocal volumes across sessions and modalities, with learned descriptors.

NeuroscienceComputer Vision

Microglia Segmentation and Tracking in Two-Photon Microscopy

Following the brain's immune cells through time — their motility, and what they touch.

NeuroscienceDeep Learning

Facial Kinematics as a Correlate of Behaviour and Neural Activity

Reading a head-fixed mouse's face — optical flow plus landmarks, against reward and neural signals.

NeuroscienceComputer Vision

3D Scene Graph Filtering using Graph Neural Networks

M.Sc. thesis — pruning 3D scene graphs against a user query before an LLM ever sees them.

Deep LearningComputer Vision

Single Shot RADAR Target Detection and Localization

A CNN-RNN detector that reads range-doppler maps in one pass — and beats CFAR.

Deep LearningRADAR

VR & Acquisition Tools for Neuroscience

Closed-loop VR for head-fixed mice, synchronised to imaging and ephys recording.

NeuroscienceEmbedded

Multi-Modular Tracking using LSTM Networks

Replacing the filter, prediction and association stages of a tracker with learned modules.

Deep LearningRADAR

Robot Air Hockey Challenge 2023

MPC and RL driving a 7-DOF KUKA iiwa14 through a game with no time to think.

Reinforcement LearningRobotics

FreiCAR: Practical Autonomous Driving

Perception and control for an RC car that has to finish the lap on its own.

RoboticsComputer Vision

Voice-Activated Vending Machine

Replacing the food-truck front desk — takes the order, checks allergies, remembers you.

EmbeddedComputer Vision

RL Frequency Hopping Against Jammers

Learning where to hop next, under timing and power constraints.

Reinforcement LearningEmbedded

Quadcopter with Auto-Levelling

PID and a complementary filter — the flight controller, written from scratch.

EmbeddedRobotics

Drag, scroll or shift-scroll to browse.

Achievements

  • 1st Place — Tree Segmentation Challenge, DL'22 Competition
  • 3rd Place — Large Networks Track, DL'22 Competition
  • Winner (Automation & Optimizations) — Voice AI Hardware Challenge, Hackster.io

Publications

  1. T. M. Dhipu, R. Verma and R. Rajesh, "Identification Friend or Foe Mode Code Detection Using Deep Pulse Detector Network," Journal of Aerospace Information Systems, vol. 20, no. 1, pp. 17–24, 2023.
  2. R. Verma, R. Rajesh and M. S. Easwaran, "Modular Multitarget Tracking Using Long Short-Term Memory Networks," Journal of Aerospace Information Systems, vol. 18, no. 10, pp. 751–754, 2021.
  3. T. M. Dhipu, R. Verma, R. Rajesh and S. Varughese, "Single Shot Radar Target Detection and Localization using Deep Neural Network," IEEE CONECCT, Bangalore, India, 2022, pp. 1–9, doi: 10.1109/CONECCT55679.2022.9865801.
  4. R. Verma, S. J. Darak, V. Tikkiwal, H. Joshi and R. Kumar, "Countermeasures Against Jamming Attack in Sensor Networks with Timing and Power Constraints," 11th International Conference on Communication Systems & Networks (COMSNETS 2019: Poster), India, Jan. 2019.