Tushar Kaushik

Software Engineer

Bengaluru, Karnataka, India9 yrs 5 mos experience
Highly StableAI Enabled

Key Highlights

  • 5 years of experience in embedded systems and AI.
  • Proven track record in leading projects and mentoring.
  • Expertise in optimizing deep learning models for edge devices.
Stackforce AI infers this person is a skilled Embedded Systems Engineer with expertise in AI and avionics.

Contact

Skills

Core Skills

Edge AiTensorflowWireless TechnologiesCAvionicsC++Graphical User Interface (gui)PythonDockerControl Systems DesignKalman Filtering

Other Skills

DebuggingPython (Programming Language)C (Programming Language)C++ 17AutopilotQt CreatorMachine LearningReal-Time Operating Systems (RTOS)RTOSInternet Protocol Suite (TCP/IP)Robot Operating System (ROS)DatabasesREST APIsPower ElectronicsAttitude Determination and Control

About

I'm a software developer with 5 years of experience in embedded systems, concurrency, device driver development, modern C++ standards, distributed systems, data structures and algorithms. I have proven track record in leading projects, mentoring team members, and enhancing performance, scalability, and cost-efficiency, all while maintaining exceptional code and documentation quality My current interest is Edge AI : running deep learning models in extremely resource constrained environments I follow top experts in the fields of my interest and try to bring a touch of novelty while utilizing the best of the existing methods.

Experience

9 yrs 5 mos
Total Experience
1 yr 7 mos
Average Tenure
1 yr 3 mos
Current Experience

Google

Software Engineer

Feb 2025Present · 1 yr 3 mos · Bengaluru, Karnataka, India

Synaptics incorporated

Software Engineer

Sep 2021Feb 2025 · 3 yrs 5 mos · India

  • Implementing and optimizing deep learning models on ARM Cortex M based edge devices, utilizing TensorFlow Lite, micro NPUs and inhouse ML compiler
  • Implemented new features, proposed changes in the standards and maintained the MAC and Network Layer of DECT wireless stack (an ETSI Standard implemented in C). These features include :
  • Automatic Dynamic Power Control : a novel master-slave algorithm to set the TX power of two connected nodes based on RSSI and link quality parameters to dynamically use less power when nodes are close and more power when nodes are far.
  • RF Ping : an extremely lightweight replacement to existing ping-pong feature in our DECT implementation
  • Paging burst : for ULE devices that only wake up once a while,
  • Emergency link setup and location triangulation using RSSI
  • The wireless stack runs on coin-cell capable Cortex M33 based chip, with lots of power and performance optimizations.
DebuggingWireless TechnologiesPython (Programming Language)C (Programming Language)C++ 17Edge AI+1

Newspace research and technologies

Design Engineer

Nov 2020Sep 2021 · 10 mos · India

  • Joined at early stage and got to work on a unique fixed wing Solar Powered UAV
  • Designed, developed (in C++) and tested a mission-critical RTOS based Data Acquisition System, running on ARM Cortex A, that diagnoses and commands the on-board avionics. Created custom protocols for interaction between various CAN (battery management systems), UART (autopilot) and wireless-ly connected nodes (ground station) that constituted the on board avionics system. This DAQ could remotely be connected to multiple end-points on the ground-station.
  • I also designed and developed, using QT Designer and Python, a multi-tab GUI that processes real-time data received from DAQ and displays it for intelligent system health monitoring and control.
  • Implemented custom applications over C++ based open source flight management stack (PX4).
AutopilotQt CreatorMachine LearningC++AvionicsGraphical User Interface (GUI)+1

Greyorange

Embedded Engineer

Jun 2019Nov 2020 · 1 yr 5 mos · Gurugram, Haryana, India

  • Designed and developed a linux based microservice using C++, ROS, REST APIs and Docker, facilitating diagnostics and control of embedded targets connected over CAN and LIN bus
  • Developed a feature rich bootloader and companion python based server application for ARM Cortex-M and Cortex-A products supporting dual-application mode, corrupt-application detection and user authentication
  • Profiled, analyzed and ported bare-metal legacy code running on ARM Cortex M devices to propriety FreeRTOS based framework utilizing modern design paradigms
C++Internet Protocol Suite (TCP/IP)Python (Programming Language)C (Programming Language)Robot Operating System (ROS)Databases+2

Tejas networks

Practice School II

Jul 2018Dec 2018 · 5 mos · Bangaluru

  • Simulated (in PSIM) and implemented a digital control scheme for ‘Phase Shifted Full Bridge DC-DC Converter’ (48V-12V stepdown) on TI Piccolo C2000, effectively replacing analog switching logic.
Power ElectronicsC (Programming Language)Control Systems Design

Isro satellite centre

Research Intern

Jun 2017Jul 2017 · 1 mo · Bangalore, India

  • Worked on ‘Relative Attitude and Position Estimation of LEO Spacecraft in Formation using Line of Sight Measurements’.
  • Implemented, on the on-board-computer, a full-state Kalman Filter and Controller in state-space without directly using target satellite dynamics data.
MatlabC++On Board Computer

Team anant

Team Lead, ADCS

Dec 2015Dec 2017 · 2 yrs · BITS, Pilani

  • Led a team of 5 members that devised, simulated and implemented orientation estimation and control algorithms like Extended Kalman Filter and sliding mode controller for de-tumbling, target-pointing and ground station tracking for a 3U hyperspectral-imaging nanosatellite
  • Selected and interfaced actuators and sensors, ported algorithm implementations to the on-board computer and built test-beds in collaboration with other subsystem teams
Kalman filteringAttitude Determination and ControlC++

Education

Georgia Institute of Technology

Master of Science - Computer Science (OMSCS) | Part Time

Jan 2023Jan 2026

Birla Institute of Technology and Science, Pilani

Bachelor’s Degree (Honors) — Electronics and Instrumentation

Jan 2019Present

Stackforce found 100+ more professionals with Edge Ai & Tensorflow

Explore similar profiles based on matching skills and experience