Chris Thaliyath

Robotics software engineer with 8+ years of experience in perception, motion planning, and autonomous systems across ADAS (Mercedes-Benz), autonomous lawn mowers (Electric Sheep Robotics), autonomous golf carts (Bosch), and AV logistics (K2.world). Expert in multi-modal perception (camera, LiDAR, radar), SLAM, sensor fusion, 6-DOF pose estimation, and real-time distributed systems. Proficient in Python, C++, PyTorch, ROS/ROS2, and simulation platforms including CARLA, Gazebo, Unity, and Isaac Sim.

Driven by continuous reinvention, human-centered technology, and values of peace, empathy, and excellence.

Autonomous Drive Test Engineer at K2.world – Kintsugi NeoLixGo (2026-05 – Present)

• Validate and verify autonomous driving stacks for last-mile logistics platforms operating in complex urban environments.
• Design and execute structured test plans covering perception, planning, and control modules across simulation and real-world scenarios.
• Collaborate with cross-functional teams to identify failure modes, reproduce edge cases, and drive resolution through the development cycle.
• Develop automated test harnesses and data-driven evaluation pipelines to accelerate regression testing and safety validation.

Lead Engineer – Autonomy at Rosh.ai (2025-10 – 2026-04)

• Team Leadership: Led a multi-disciplinary team of 10+ engineers; oversaw end-to-end development of autonomous vehicle platforms and industrial mobile robotics.
• Fleet Conversion: Architected the autonomy stack for converting Suzuki Ertiga and Tata Punch platforms into autonomous shuttles, managing the full transition to drive-by-wire.
• Brownfield Retrofit: Designed modular hardware and embedded software frameworks for retrofitting legacy vehicle systems with modern autonomous capabilities.
• Trucking & Parking: Technical lead for L2/L3 features in autonomous trucking and a hybrid USS + Fisheye based auto-parking solution for passenger vehicles.

On-Site Deployment — Rivian Motors, Bloomington, IL (USA) at Rosh.ai (2026-02 – 2026-03)

Led the on-site deployment of an autonomous Auto-PDI (Pre-Delivery Inspection) Robot at Rivian Motors' manufacturing facility. Conducted end-to-end quality inspections for End-of-Line (EoL), HVAC, Noise/Vibration/Harshness (NVH), and steering/pedal returnability KPIs — directly integrating the autonomy stack into a live production environment.

Level 3 ADAS – HAF 95 & HAF 135 at Mercedes-Benz Research and Development India (2022-08 – 2025-10)

• Architected production-ready real-time distributed control modules for L3 ADAS, defining system boundaries, interfaces, and requirements.
• Developed LiDAR-based clustering, tracking, and classification pipelines using Euclidean clustering, Hungarian matching, CenterPoint3D, and PointNet for small object detection.
• Implemented monocular depth estimation and BEV-based perception models (PyTorch, BEVDet) to replace LiDAR with camera-only occupancy grids.
• Hands-on with multi-modal BEVFusion (Camera + LiDAR); explored Vision-Language Models (VLMs) for semantic scene understanding.
• Built multi-LiDAR and multi-camera calibration pipelines using ARUCO markers; implemented ICP/NDT-based alignment and ego-motion estimation.
• Designed SLAM–IMU–GPS fusion achieving centimeter-level localization accuracy.
• Implemented MPC, LQR, Pure Pursuit, and 2-DoF controllers integrated with vehicle dynamics.
• Maintained Gen5 ADAS validation pipelines using Airflow, Docker, CI/CD, and large-scale data replay.
• Deployed perception and planning stacks on NVIDIA Orin (IDC6) optimized with TensorRT and CUDA.
• Developed CAN/CAN-FD drivers, DBC decoding, and ROS interfaces.

Fully Autonomous Lawn Mowers at Electric Sheep Robotics (2021-07 – 2022-08)

• Improved SLAM-based 6-DOF pose estimation using EKF fusion of LiDAR, camera, IMU, and GPS.
• Built navigation pipelines across Isaac Sim, Gazebo, and Unity enabling sim-to-real transfer.
• Designed synthetic data pipelines for Vision-Language Models.
• Integrated Visual Teach & Repeat (VT&R) navigation (Under the guidance of Prof. Tim Barfoot, University of Toronto).
• Deployed autonomy stack on NVIDIA Orin.

Spatial Computing & AR Platforms at Arway.ai (2020-01 – 2020-12)

• Implemented kidnapped-robot relocalization using Bag-of-Words techniques.
• Cross-compiled ORB-SLAM2 C++ binaries for Unity, Android, and iOS.
• Built a custom SLAM-based relocalization system replacing ARKit/ARCore.

Senior Software Engineer (Promoted from Software Engineer) at Bosch Global Software Technologies (2020-01 – 2021-08)

• High-Definition Mapping: Engineered a full-pipeline solution for campus-scale navigation by generating Lanelet2 and OpenStreetMap (OSM) compliant maps, enabling lane-level semantic context for autonomous routing.
• Localization Stack: Developed and deployed a robust LiDAR-SLAM system utilizing Velodyne VLP-16; implemented multi-sensor EKF fusion (LiDAR + GPS + IMU) to maintain centimeter-level drift-free localization in GPS-denied environments.
• Extrinsic Calibration: Designed and implemented automated calibration routines for LiDAR-to-LiDAR and LiDAR-to-Camera setups, ensuring precise point-cloud projection and sensor data alignment for downstream perception.
• Planning & Control: Built waypoint followers and integrated lane-keep assist (LKA) and adaptive cruise control (ACC) logic tailored for low-speed autonomous golf carts.
• Hardware Integration: Developed custom ROS drivers for vehicle interfaces (CAN, Electronic Power Steering, throttle/brake) and environmental sensors (Radar, Ultrasonic).
• Embedded Deployment: Optimized and deployed the full autonomy stack on NVIDIA Drive PX platforms operating on RTOS (QNX and Linux-RT) to meet safety-critical timing requirements.

Software Engineer – Autonomous Low-Speed Golf Cart at Bosch Global Software Technologies (2018-09 – 2020-12)

• Drive-by-Wire (DBW) Development: Developed and integrated ROS-compatible drivers for vehicle actuation systems including EPAS (Electronic Power Assisted Steering), electronic throttle, and braking systems.
• Sensor Integration: Implemented hardware abstraction layers and drivers for a diverse sensor suite: Ultrasonic Sensors (USS), Mid-Range Radar (MRR), and Wheel Speed Sensors for odometry.
• Communication Protocols: Developed CAN/CAN-FD interfaces and DBC decoding logic to bridge the gap between low-level vehicle controllers and the high-level ROS navigation stack.
• Control Theory: Implemented basic waypoint following and PID-based longitudinal/lateral control integrated with vehicle dynamics.

Agricultural Autonomous Robot – Rubber Tapping at Vanora Robots (2017-06 – 2018-03)

• Developed ROS drivers for Sabertooth 2x32 motor controllers on Raspberry Pi.
• Programmed embedded controllers (Arduino, STM32) for real-time motor control.
• Implemented encoder-based odometry integrated with ROS navigation stack.
• Built Visual SLAM pipelines using depth camera (Xbox Kinect) fused with odometry.
• Designed ArUco marker–based auto-docking for wireless charging.

M.Tech in Artificial Intelligence & Machine Learning – Indian Institute of Technology Jammu (2023-01 – 2025-12)

B.Tech in Computer Science – Shiv Nadar University (2013-01 – 2017-12)