Siddharth Rohit Jain
A Master's student in Robotics and Autonomous Systems, focusing on Bio-Inspired Robotics. Expertise in embedded systems, ROS, and AI pipeline development. I am adept at crafting innovative solutions for complex challenges in robotics, AI, and autonomous systems.
Embedded Systems, Robotics & Controls Enthusiast
- Highest Education: Master of Science
- Majors: Robotics and Autonomous Systems
- City: Tempe, AZ
- Email: [email protected]
Expertise in embedded design using NRF BLE, ESP32, and Raspberry Pi,
emphasizing real-time control and integration.
Versed in software tools like Docker and Arduino IDE, and hardware design
with Altium. Adept in software-hardware synergy for control systems.
Strong foundation in protocols such as SPI, I2C, CAN Bus, and RF integrations
like LoRa and BLE for optimal communication in robotics.
Skilled in cloud integration for embedded systems using AWS services,
facilitating scalable and efficient control & robotic solutions.
Embedded Systems
Robotics and Controls
Machine Learning
AWS
Experience
Enterprise Technology
October 2022 - Present
Embedded Systems Engineer
- Implemented AES Encryption in a UHF mesh protocol for enhanced security in ASU cart tracking, utilizing technologies such as C++, LoRa, MQTT, and AWS IoT Core.
- Developed a BLE mesh network on ESP32 for efficient SOS alerts, integrated with React-Native for seamless operation on Android and iOS platforms, thus improving emergency response capabilities.
- Orchestrated the creation of AWS Lambda functions in conjunction with API Gateways and Timestream, resulting in a 30% improvement in data accuracy and retrieval speed, achieved through proficient use of Python. Also optimized the mpu9250 sensor in IoT trackers, significantly extending battery life to 3 years and reducing maintenance costs by 40%.
Enterprise Technology
July 2023 - Present
ML Ops and AI Development Engineer
- Developed a Python library for accessing 45 LLM models, routing between vision and audio using Lambda and API Gateway.
- Deployed Terraform plans to create and manage AWS clusters and instances for POC, beta, and production stages.
- Managed SQS for Lambdas via Terraform and deployed multi-stage Dockers on Kaniko for continuous deployment.
- Architected LLM deployment on AWS Lambda using binaries in C, achieving 10 tokens/sec, optimizing cost-efficiency.
- Integrated text-to-speech, speech-to-text, and facial recognition on a robot head running on Raspberry Pi using multi-threading.
- Added Lambda Insights for memory profiling, effectively reducing the memory required for every lambda and saving cost.
Enterprise Technology
October 2023 - July 2023
Embedded Systems Engineer Engineer
- Implemented AES-128 encryption to enhance the security of custom UHF mesh networks using MQTT on an edge device.
- Engineered a BLE LoRa mesh network on ESP32 for SOS alerts, significantly improving emergency response efficiency.
- Optimized MPU9250 sensor in IoT trackers, extending battery life to 1 year by enabling deep sleep mode.
- Maintained a LoRa and LoRaWAN mesh network with 25 nodes to track golf carts on campus via MQTT on AWS.
Bio-Inspired Robotics, Technology and Healthcare Lab (BIRTH Lab)
Dec 2022 - Present
Research Assistant
- Automated friction analysis of PDMS pads on curved surfaces, leading 180 experiments for thesis research.
- Designed a 3-axis testing apparatus with a 6-axis load cell and a PID controller, achieving robust control.
- Controlled UR-16e 6-axis robotic arm for load-carrying tasks and designed a SpaceMouse controller in ROS.
DJS Kronos India
Mar 2019 - May 2021
Vice Captain
- Led the design of a 4WD ATV on Simulink, achieving a 17% increase in operational efficiency; awarded 2nd Best 4WD Team.
- Built a DAQ system using the GSM SIM 900 module on a Raspberry Pi Zero via ThingSpeak Communication.
Projects
Dexterous Manipulation with a Robotic Hand
AWAC, Reinforcement Learning
- Developed an Advantage Weighted Actor-Critic algorithm to enhance the performance of a 6-DoF robotic hand. Achieved up to a 20% improvement in dexterous manipulation success rates.
Multi Robot Search and Rescue
Gazebo, ROS, RVIZ, RTAB
- Developed a decentralized quadcopter swarm with Potential-Field and Frontier Exploration algorithms for dynamic 3D mapping.
- Enhanced exploration efficiency with a leader-trooper strategy, navigating through complex terrains and avoiding local minima.
- Conducted ROS and Gazebo simulations, validating the swarm's ability to produce detailed maps and critical visual data.
Self Balancing Stewart Platform
PID Controller, MATLAB, Simulink
- Masterfully engineered a closed-loop PID controller for the Stewart platform using Simulink, strategically optimizing ball motion stability.
- Achieved a breakthrough in control precision by reducing Steady State Error through integral tuning, enhancing system responsiveness to an impressive 0.5 seconds, reflecting a seamless synergy of theoretical control principles and hands-on implementation.
UAV Line Follower Drone
MATLAB, Simulink, Edge Detection
- Pioneered a cutting-edge Line Follower function for the Parrot Mambo Mini-Drone, harnessing advanced edge detection techniques to rapidly calculate the nearest edge.
- Skillfully identified specific HSV values of the track within an astounding 20 ms using Simulink, culminating in a 95\% accuracy rate across 40 rigorous tests, marking a significant milestone in UAV navigation and control.
Statistical Analysis of Machine Learning Algorithms for Fraud Detection
Deep Learning, RNN, One-Hot Encoding
- Masterminded a robust fraud detection system by meticulously analyzing and employing cutting-edge machine learning algorithms on an expansive dataset of 1 million bank transactions. Leveraged one-hot encoding, TensorFlow, and innovative preprocessing strategies like SMOTE in Python to optimize model training.
- Skillfully validated model variations through a rigorous feature importance assessment and precise statistical evaluations, resulting in a remarkable 97.2\% accuracy rate, setting a new benchmark in the field of financial security.
NaviGate Console Application
C Programming, Console Application, File Navigation
- Designed and developed NaviGate, a user-friendly console application for navigating directories and viewing file contents on Windows systems using a command-line interface. Developed in C, it highlights my expertise in programming and system development.
Custom LoRa and Ethernet Communication Board
ESP32 S3, PCB Design, Embedded C, FreeRTOS
- Conceptualized and masterfully developed a state-of-the-art communication board, integrating Xtensa LX7 (32-bit, 240 MHz), RFM95W LoRa (915 MHz, 15 km range), and LAN8720 Ethernet, establishing a new benchmark in versatile wireless and wired connectivity.
- Executed a meticulous 4-layer PCB design with 50-ohm impedance control for impeccable RF signal integrity, leveraging innovative techniques in embedded C and FreeRTOS.
- Crafted concurrent communication tasks, OTA updates, and advanced power management strategies that pushed the envelope of efficiency and reliability.
- The resultant system not only stands as a testament to cutting-edge engineering but also serves as a blueprint for future advancements in the integration of LoRa and Ethernet technologies.
- Conducted comprehensive testing for performance validation, including RF signal quality analysis, Ethernet throughput benchmarking, and system latency measurement.
Dynamic Pathfinding in Complex Environments
Python, Matplotlib, Algorithm Design, Dynamic Programming
- Innovated in the field of dynamic pathfinding by skillfully developing, analyzing, and comparing a series of cutting-edge algorithms such as A*, Dijkstra's, DFS, BFS, and Greedy Best-First Search.
- Utilized Python and Matplotlib to adapt these algorithms to complex, real-world scenarios with moving obstacles, considering metrics like path length (24-25 steps) and time (0.0011 to 3.288 seconds).
- The novel approach led to unparalleled efficiency in navigation, providing a visually engaging analysis that illuminated new insights into the behavior of each algorithm, setting a precedent in adaptability and responsiveness in pathfinding solutions.
Patents & Publications
Steering Knuckle Joint for Double A-arm Suspension System
Patent No. 378832-001
Four-wheel Drive, r-zeppa Joint, All-terrain Vehicle Design
- The steering knuckle joint is an integral component that links the suspension arms, steering tie rod, and wheel assembly.
- Specifically designed for the double A-arm suspension system of a four-wheel-drive all-terrain vehicle, this knuckle incorporates the r-zeppa joint as an active spindle. Shocks are strategically positioned on the upper A-arm.
- The knuckle also boasts an inbuilt steering arm, directly connecting to the tie rod, ensuring seamless linkage.
Design Of Single Stage Open Differential (Pending)
Patent No. 378831-001
Automobile Dynamics, Differential Mechanism, Power Distribution
- The open differential is a pivotal mechanism in automobiles, especially when navigating turns. Due to the differing distances traveled by the inside and outside wheels, there's a need for them to rotate at unique speeds while still being driven by the engine.
- This is achieved using the differential. Apart from ensuring smooth turns, it serves to reduce the rotations from the gearbox before being transmitted to the rear axles.
- This device also changes the power's axis of rotation by 90° and efficiently distributes power to the rear driving axles, whether the vehicle is moving straight or turning.