I'm Adithya Kambat Shankar, a Machine Learning and Robotics Engineer with a deep passion for pushing the boundaries of technology. Graduated with MSc in Electrical Engineering at Pennsylvania State University, I bring a strong background in AI and Machine Learning.
I specialize in developing innovative solutions that blend machine learning and embedded systems. My work is driven by a passion for leveraging AI to solve complex problems and enhance everyday life through automation and smart technologies.
- Machine Learning & AI: Achieved 99% accuracy in multi-class classification tasks by developing and deploying advanced neural networks on embedded platforms, including Nvidia Jetson. Implemented CUDA-accelerated computer vision tasks, boosting performance 3x on edge computing devices.
- Computer Vision: Expertise in real-time image processing, implementing Semantic Segmentation, Object Detection, and Image Classification on embedded platforms to enable fast, accurate decision-making. Led a project involving Nvidia Jetson, achieving significant performance and efficiency gains in Semantic Segmentation.
- Embedded Systems & IoT: Designed and optimized autonomous systems, reducing hardware costs and improving efficiency by integrating SLAM for robots and leveraging microcontrollers like Arduino and Raspberry Pi. Developed a low-cost home surveillance system using wireless sensor networks (WSNs), cutting costs by 50% through strategic use of affordable components. Engineered a lightweight drone platform, reducing take-off weight by 300% and improving power economy using advanced PCB design and custom carbon fibre materials.
- Cloud Computing: Hands-on experience with AWS, leveraging EC2, S3, and other cloud services to create scalable and efficient data processing pipelines.
- Programming Languages: Python, C, C++, Matlab, Rust, R
- AI & Data Science: PyTorch, Keras, Scikit-learn, TensorFlow, HuggingFace, LangChain, Numpy
- Embedded & Robotics: ROS, SLAM, Arduino, Raspberry Pi, Nvidia Jetson, ESP8266/ESP32, STM32
- Cloud Platforms: AWS (S3, EC2, Bedrock), GCP, PostGres,SQLite
- Others: OpenCV, Conda, Git, MATLAB
- Graphene-based ISFET Chemisensor: Developed an Artificial Neural Network (ANN) with Tensorflow, achieving 99% accuracy in multi-class classification, a breakthrough in advanced chemisensor systems.
- Autonomous Drone Development: Extended drone flight time by 30% through EDGE computing and autonomy enhancements using Nvidia Jetson Nano and on-board NDVI and Semantic Segmentation.
- Cost-efficient Home Surveillance: Reduced the cost of home surveillance systems by 50% using a combination of affordable hardware components and computer vision technologies.
- Published Research: Contributed to the field with publications in IEEE and Springer, including works on smart aquaculture management systems and crop health monitoring using UAVs.
- AI-Powered Traffic Analysis: Developing a privacy-preserving traffic detection system with over 90% accuracy, ensuring zero identifiable signatures.
- Reinforcement Learning in Gaming: Training AI models using PPO algorithms to surpass human-level performance in complex gaming environments like Rocket League.
- Cloud-Integrated IoT Solutions: Building IoT frameworks that leverage cloud computing for enhanced performance and real-time data processing.
- Stock trading with ARIMA models: Building an ARIMA model to help predict stock price trends. Invested in stocks without outliers and got a YoY return of 25% on average.
- AI-Powered Traffic Analysis: Developing a privacy-preserving traffic detection system with over 90% accuracy, ensuring zero identifiable signatures, making improvements to the simulation models and the detection algorithms.
- AI for Healthcare: Training a multi-modal AI solution for improving patient outcomes, leveraging multiple models to build a cohesive system. By utilizing different models for each modality and aggregating their outputs into a unified system, healthcare providers can offer more personalized and data-driven treatment plans. Can implement with explainable AI to improve the models.
- Neural Networks on Rust: Building projects on microcontrollers and using rust to develop some parallels to the projects I am working on.
I'm always on the lookout for exciting opportunities to collaborate on groundbreaking projects in AI, IoT, and Robotics. If you're working on something innovative and think we could create something amazing together, feel free to reach out!
- Email: adithya7shankar@gmail.com
- LinkedIn: linkedin.com/in/adithya7shankar
When I'm not deep in code or driving my next groundbreaking project, you'll likely find me exploring cutting-edge advancements in AI or getting hands-on with DIY electronics and robotics kits. And if you catch me on Steam, I'm probably strategizing in Dota 2 or scoring goals in Rocket League with friends. I'm driven by a passion for continuous learning and thrive on tackling new challenges head-on! Check out my repositories and see the work I’ve been doing! 🚀