United States

Technical Product Management

Program Management

Hardware Design

AI/ML Engineering

Soham Banerjee

Senior Embedded Software Engineer

FreeCodeCamp Articles

Collaboration

Problem-solving

Empathy

Team player

Organisation skills

Leadership

Analytical skills

Machine learning

Unit Testing

Pandas

Python

GitHub

Matlab

Confluence

Numpy

- Technical Lead for Audio Products Battery, Power and Charging Algorithms.
- Design and architect the deployment of ML models on 32-bit processors and create infrastructure for ML model validation
- Define system architecture and design new algorithms to enhance customer experience on audio products
- Design, Implement, test and debug embedded C/C++ code running on MCU based circuits for Apple audio products
- Collaborate with software, testing and HW teams to define feature specs
- Use python and scripting languages to design automation tools for testing features

My responsibilities span all aspects of product development: 
- Collaborate with marketing and cross-functional engineering teams to define product features and requirements
- Play a lead role in defining system architecture to implement new audio products and features
- Design, implement, debug, and test embedded C and C++ code running on 8-bit, 16-bit, and 32-bit MCU's for Apple-branded audio products
- Support electrical bring-up and debug of prototype hardware
- Work with internal teams and external contract manufacturers to bring products to mass production
- Collaborate with software teams to define communication protocols
- Support audio products through test scripts and tools

Team: 
I was working in the Audio Accessories Firmware Team at Apple Inc. This team is responsible for design and development of firmware for audio accessories products at Apple.

Responsibility:
I was given the responsibility to design a gas gauge firmware for the Airpod battery and the Airpod case battery. 
Initially, I designed a battery characterizer script in Python, the primary purpose of the script was to extract useful data out of excel sheets given by the battery team. Additionally, battery protection algorithms were implemented in the script to maximize shelf life The output consisted of data sets used for charge and discharge cycle of the battery and other relevant parameters to be used during SOC estimation for a range of temperatures and loading conditions. This script could be ported to other projects with slight changes in the parsing script. This script was initially designed for the Airpod battery and later ported to Airpod case.
The main objectives of the firmware were State of Charge(SOC) estimation of the battery and scaling them to (0 - 100%), implementation of battery protection and low power operation, and control functionalities related to the battery charger. State of charge of the battery was estimated using a temperature based interpolation technique using a moving average algorithm to filter out noise from the input. 
Additionally, I was involved in power analysis and workflow testing of the Airpod and Airpod case and in writing python scripts for parsing data logged from wireless testing and output useful data for analysis.
It was a great learning experience at Apple and it helped me enhance my technical knowledge as well as knowledge regarding system development life cycle of a product at Apple.

Assistant to Professor Timothy Lee Scherr for the course on Mastering Embedded Systems Architecture, (ECEN 5003) and Real-Time Embedded Systems (ECEN 5623) under the Embedded Systems Program for Spring 2016. 

My tasks as TA include 
- course and project planning
- designing of project guides
- assisting students regarding course material and holding office hours.
- conducting interview and live demo for projects
- grading of quizzes and projects.

Team:
I am a research assistant at the Autonomous Robotics and Perception Group under Professor Christoffer Heckman. I am currently working in the autonomous car project also called the Ninja Car project. I am primarily responsible for the firmware design of the car locomotion control unit with Sina Aghli, a Phd Student in the Department of Computer Science. 
I also worked in an independent study project under Professor Rick Han and with Professor Christoffer Heckman on a project which aims to develop a solution for dynamic obstacle avoidance using RTOS. 

Responsibility: 
- Development of Firmware for the locomotion unit in the Ninja car platform to control a servo motor and a 3-phase BLDC motor. The locomotion unit uses an ARM Cortex-M3 at its core to control the servo and the BLDC driver which drives the hex-bridge unit. The locomotion unit also interfaces an onboard 9 axis IMU sensor for posture estimation. A DRV8302 is used as a BLDC driver to drive the MOSFETs of the hex-bridge unit. The code is mostly written using STM32CUBEMX Hardware Abstraction Layer. The firmware is currently being ported to FreeRTOS to take introduce deadlines for various functions and making the whole system deterministic.

- Development of a solution for dynamic object avoidance for a robot using real time operating system. The purpose of the real time operating system is to identify both moving and stationary obstacles and take immediate action by running an obstacle avoidance algorithm that controls the motors of the robot with the help of the camera sensor. All normal tasks such as perception, planning and control are bypassed. I compiled a Xenomai real-time kernel on a linux kernel as a platform for this project. The choice of Xenomai over Rtai and Real-time patch was made due to the ability to support a variety of platforms and also to prevent a large number of context switch and overall less overhead because of its dual kernel approach.

Team:
I was working in the Infotronics and Advanced Electronics team under the Advanced Technology sector of the Auto & Farm Division in Mahindra and Mahindra. This group works in various kinds of technology projects relevant to automobiles of all segments like commercial, personal and farm. 

Responsibility:
- Worked in the Connected Vehicle Solutions project which aims to connect the automobiles under various segments to Mahindra customer care team, to customers like individual owner or fleet owner, to dealers and to service centers. 
- Responsible for the design, testing and Failure Mode Effect Analysis of the hardware device(CCU) which was fitted in the vehicles to transmit data from the vehicle to the Mahindra cloud, data was either captured from the sensors which were hard-wired to the CCU or from the CAN network. The CCU was made tamper proof to increase the safety and authenticity of the data.  
- Performed an IFMEA (Interface level, FMEA) of the CCU for a commercial segment vehicle to come up with a Design Verification Plan for the CCU fitted on to the vehicle. During the development of the IFMEA document, I got to analyze and understand the system in a better way and identify most of the failures that may lead to the system malfunctioning. It was a challenge to reduce the Risk Priority Number for all the failures. 
- Also responsible in the design and workflow of the User Interface for the User management and Vehicle Management section of the Webpage under Fleet Management and Emergency Services. The design followed a modular approach to keep things easy at the designer’s end for future modifications and additions.

Team:
I was working in the motorcycle team under the EEE department for the implementation of prototypes on new concepts pertaining to two wheelers and was also involved in one of the mainstream motorcycle projects 

Mainstream Project (Mojo):
 Worked on a mainstream project which was based on a 300cc Sports motorcycIe named MOJO which has been launched in the market in 2015.

Responsibility:
- Tested and validated the hardware components of the EFI system (Electronic Fuel Injection) on the motorcycle as well as conducted tests to test and improve important parameters such as mileage and drive-ability of the motorcycle. The tests conducted on the EFI components mostly consisted of functional tests, overvoltage short to battery and short to ground tests performed on the motorcycle as well as bench rigs. 
- Calibrated performance parameters and actuator timings of the EFI system to increase overall performance of the motorcycle. Designed various safety features for prevention of accident hazards on the motorcycle like the use of roll-over sensors. 
- Tested and validated the charging system which involved the magneto (3 phase synchronous alternator), the regulator rectifier unit and the battery. The tests involved the use of series and shunt regulator rectifiers units and three phase star configuration synchronous alternators and lead-acid VRLA batteries. Variable load tests and endurance tests were performed to ensure reliability and robustness of the system.

Under the technology projects, my responsibility was to design a working prototype on a new concept pertaining to two wheelers and also perform a failure mode effect analysis (FMEA) for the method chosen to implement the concept. The projects are added under the Projects Section below.

Soham Banerjee🇺🇸

About Soham Banerjee

Expertise

Sessions with Soham Banerjee

What mentees say