Meal planning for iOS.
React Native application for meal planning, grocery lists, and food journaling, with persistent local state and optional AI meal generation.
Kshitij
Prasad
Robotics & software engineer.
I develop vehicle-data pipelines and ADAS scenario retrieval at Audi. Previously at Akkodis, I worked on ROS 2 integration and reinforcement-learning simulation.
INGOLSTADT, GERMANYPROJECTS ↓
01 / SELECTED PROJECTS
Projects
Software, robotics, and automotive data.
Insurance renewal software.
Questionnaire preparation with source references, revision history, and approval controls for commercial insurance renewals.
Reinforcement learning in Gazebo/PX4.
Benchmarked PPO, SAC and TD3 in Gazebo/PX4 SITL with trajectory logging and repeatable evaluation. PPO achieved 5.5 cm mean final horizontal error over 100 simulation episodes.
Telemetry workflows for ADAS simulation.
Vehicle telemetry pipelines, computer vision, and CLIP semantic search for retrieving and reconstructing real driving scenarios.
Fleet planning and traffic visualisation.
An interactive fleet-planning prototype with simulated demand and an optional PPO backend. First place at the PAVE Europe Next Gen Challenge in Brussels.
Software for a driverless racecar.
Real-time perception, motion planning, and sensor integration in C++ / ROS 2 on NVIDIA Jetson for an autonomous Formula Student racecar.
FURTHER PROJECTS
07Akkodis service robotROS 2 · Nav2 · RealSense · Object detectionROBOTICS
Connecting perception, object detection, and navigation for a physical service robot.
08Aero-XRLAirbus Fly Your Ideas · Explainable AIAI PRODUCT
A decision-support prototype combining explainable AI, data pipelines, and dashboards for aviation.
09Pixhawk UAV platformPX4 · MAVSDK · Embedded flight controlDRONE ENGINEERING
A Pixhawk-based platform for hardware assembly, flight-controller integration and future policy testing.
10VR accident reconstructionDEKRA · CARLA · IPG CarMakerSIMULATION
Exploring accident scenarios through simulation, reconstruction, and immersive visualization.
11Campus HelpNext.js · TypeScript · PostgreSQLWEB PRODUCT
A peer-tutoring platform connecting students at TH Ingolstadt.
THREE.JS / INTERACTIVE VISUALISATION
UAV landing
visualisation
A Three.js illustration of a quadrotor approaching a moving platform. Pause the sequence, change the camera, or inspect a point on the timeline.
SCENARIO STATE
- Time
- Height to deck
- Target offset
Illustrative trajectory. Recorded Gazebo runs and benchmark results are available in the research project.
Read the thesis ↗02 / EXPERIENCE
Experience
MAY 2026 | PRESENTCURRENT
AUDI AG
Ingolstadt, GermanyData Engineering & AI · Driver Assistance Internship
Vehicle-data processing and scenario retrieval for driver-assistance simulation.
- Python and C++ pipelines for vehicle-bus, camera and sensor telemetry.
- CLIP-based semantic search for retrieving driving scenarios through natural-language queries.
- Computer vision, object tracking, and scenario reconstruction for ADAS evaluation across PPE and PPC vehicle platforms.
OCT 2025 | MAR 2026
AKKODIS
Ingolstadt, GermanyBachelor’s Thesis · Autonomous Aerial Systems
Developed a Gazebo/PX4 SITL landing benchmark with ROS 2 integration, policy training, trajectory logging and standardized evaluation of PPO, SAC and TD3. Physical policy transfer remains future work.
APR 2024 | OCT 2025
AKKODIS
Ingolstadt, GermanySoftware Engineering · Intern & Working Student
ROS 2 service robot integration, C++ perception and sensor interfaces, Nav2 navigation, and CARLA simulation scenarios for automated driving evaluation.
OCT 2022 | OCT 2024
Schanzer Racing e.V.
TH IngolstadtDriverless · Formula Student
Real-time motion planning, multi-sensor integration, and embedded C++ / ROS 2 software on NVIDIA Jetson for an autonomous racecar.
EDUCATION / TH INGOLSTADT
Autonomous Vehicle Engineering
B.Eng. · Graduated 2026
Bachelor’s thesis · Grade 1.0
UAV precision landing research
INGOLSTADT, DE
03 / BACKGROUND
About
I’m based in Ingolstadt, Germany, with a B.Eng. in Autonomous Vehicle Engineering from TH Ingolstadt. My experience includes automotive data processing at Audi, ROS 2 integration at Akkodis, and driverless software for Schanzer Racing.
I work primarily in C++, Python, and TypeScript. My independent projects include Nouri, a meal-planning app, and Scope, an insurance-renewal workflow.
TECHNOLOGIES
Tools and platforms
Software, simulation and hardware integration.
01 / SOFTWARE
C++ · Python · TypeScript
SQL · ROS 2 · Linux · Docker · Git
02 / MACHINE LEARNING
PyTorch · Deep RL
PPO · SAC · TD3 · CLIP · Computer vision
03 / SIMULATION
Gazebo · CARLA
PX4 SITL · Digital twins
04 / HARDWARE
Jetson · Pixhawk
LiDAR · RealSense · GNSS · IMU
BRUSSELS / 2025
PAVE Europe
Next Gen Challenge
First place at the PAVE Europe Next Gen Challenge for work on autonomous mobility and public acceptance.
Recommendation from Prof. Dr.-Ing. Michael Mecking ↗
