Extensive experience in designing and implementing various analog and digital communication systems. My expertise includes: • Modulation techniques: DSB-SC, FM, AM, ASK, FSK, PSK, QAM • Data encoding: PCM, DPCM, correlative coding • Multiplexing: Time Division Multiplexing (TDM) • Performance analysis: BER, ISI using eye diagrams • Tools: MATLAB, Python, Simulink My work in this field has involved signal processing, spectrum analysis, and performance optimization.
Significant experience in designing and analyzing CMOS circuits, implementing projects using Virtuoso Cadence. My expertise in this field includes: • CMOS Logic Gates: Analysis of CMOS inverter characteristics, noise margins, delay models (Elmore and RC Delay), and transient performance optimization. • VLSI and Fabrication Technology: Understanding of the CMOS fabrication process (n-Well, p-Well, Twin-tub processes) and recent trends like FinFET and ultra-thin body MOSFETs for scaling and addressing short-channel effects. • Combinational and Sequential Circuits: Designing combinational logic networks (Pseudo nMOS, Clocked CMOS, Dynamic CMOS) and sequential circuits such as latches and flip-flops, including clocking and distribution strategies. • Tools and Platforms: Proficient in using Virtuoso Cadence for circuit design and simulation. • Advanced Concepts: Familiarity with cutting-edge VLSI technologies and their applications in modern electronics.
Expertise in operating systems and embedded systems, including practical experience implementing concepts on LPC21xx and LPC17xx development boards. Key areas of focus: • Operating System Concepts: Process management, inter-process communication, CPU scheduling algorithms (FCFS, SJF, Priority, Round Robin), memory management techniques, and system boot structures like monolithic and microkernels. • Real-Time Operating Systems (RTOS): Familiarity with RTOS architecture, task synchronization using semaphores, event flags, message queues, and addressing priority inversion problems. • Programming with RTX Kernel: Developed task synchronization and inter-task communication programs using the RTX kernel. • Embedded C Programming: Extensive experience in writing efficient code for embedded systems. • System Design: Proficiency in designing systems for real-time and embedded applications.
Active member and lead of the KLE Rocketry Club, focusing on various space-related projects: • Satellite Development: Contributed to miniature satellite projects including CanSat and CubeSat, gaining hands-on experience in satellite systems engineering. • Rocket Design: Involved in designing and optimizing rocket structures and propulsion systems. • Electronics for Space: Developed electronic systems for rovers and other space-related applications. • Research Initiatives: Actively participating in research aimed at advancing space technology applications.
Co-authored a research paper with the KLE Rocketry Club, which was published and presented at IEEE NKcon International Conference
Currently working on an innovative space research and development startup
Project: Fusion of Machine Learning and Digital Signal Processing for Landmine Detection Using Ground Penetrating Radar (GPR) Description: This project focused on developing a robust system to detect buried landmines using Ground Penetrating Radar (GPR) by combining Digital Signal Processing (DSP) techniques with Machine Learning (ML) models. The approach involved preprocessing GPR signals to remove noise, extracting key features using wavelet transforms, and classifying anomalies with ML algorithms. Key components of the project included: • Signal Preprocessing: Applied band-pass filtering and time-zero correction for noise removal and depth alignment. • Feature Extraction: Used discrete wavelet transform to identify reflection intensity and time delay, which indicate potential landmines. • Classification: Leveraged support vector machines (SVM) for distinguishing landmines from other subsurface objects. Outcomes: • Improved signal clarity with a 20 dB Signal-to-Noise Ratio (SNR). • Achieved 86% detection accuracy with an 8% false-positive rate. • Validated the approach on real-world datasets, demonstrating its applicability in post-conflict zones. This experiment highlights the potential of integrating DSP and ML techniques for solving complex real-world challenges and ensures scalability for future field implementations.
Project: Generalized Framework for Data Fusion in Connected & Autonomous Vehicles Description: Working at the Center for Intelligence and Mobility at KLETECH, I'm part of a team developing a generalized framework for data fusion in Connected & Autonomous Vehicles (CAVs). This project aims to improve the accuracy of perception derivatives and facilitate real-time processing with reliable time synchronization. Key components of the project include: • Data Collection: Gathering data from various heterogeneous ADAS sensors and inspecting the collected data. • Data Aggregation: Consolidating the collected data into a unified, accessible format and addressing variations in sensor data formats and sampling rates. • Data Fusion: Interpreting aggregated data using advanced algorithms to improve perception accuracy and handling redundancy, discrepancies, and noise effectively to generate meaningful insights. Methodology: • Data Integration: Incorporating data from sensors like GPS (latitude/longitude), IMU (speed), LiDAR, and camera feeds, and standardizing their formats for seamless processing. • Data Fusion Approach: Utilizing techniques like Kalman filtering, Bayesian inference, and machine learning models to interpret sensor data and handle sensor redundancy and discrepancies. • Time Synchronization: Developing a custom synchronization mechanism for asynchronous sensors and implementing protocols for time alignment. Contributions: • Built a Python-based toolkit for analyzing and preparing complex datasets. • Developed systems to classify data types, clean datasets, and synchronize time-series data. • Designed a sensor fusion system to combine data from various sources like GPS, LiDAR, and radar, making it more accurate and reliable. Results: • Successfully processed and merged data from multiple sensors to create a unified view. • Achieved a real-time data processing pipeline that works without delays. • Proved the system's ability to manage and synchronize diverse sensor data effectively, making it ready for real-world scenarios like autonomous vehicles. This project demonstrates my ability to work on cutting-edge technologies and contribute to the development of autonomous vehicle systems.
Experience with various generative AI architectures and techniques, focusing on deep learning models for content generation: • Autoencoders: Implemented traditional autoencoders for dimensionality reduction and feature learning. Worked on encoder-decoder architectures to compress data into a latent space representation and then reconstruct it, with applications in image denoising and anomaly detection. • Variational Autoencoders (VAEs): Developed probabilistic models that learn the underlying distribution of data in a latent space. Implemented the reparameterization trick to enable backpropagation through stochastic nodes, allowing for controlled generation of new samples by sampling from the learned latent space distribution. • Generative Adversarial Networks (GANs): Built and trained GAN architectures consisting of: - Generator networks that learn to create realistic synthetic data from random noise - Discriminator networks that learn to distinguish between real and generated samples - Implemented various GAN training techniques to address mode collapse and convergence issues • Applications: Applied these generative models to various domains including: - Image synthesis and manipulation - Data augmentation for improving model robustness - Style transfer between different domains - Anomaly detection in sensor data for autonomous systems • Tools and Frameworks: Utilized PyTorch and TensorFlow for implementing these architectures, with experience in hyperparameter tuning and optimization techniques specific to generative models. This experience with generative AI complements my work in autonomous systems and signal processing, providing additional tools for data generation, augmentation, and analysis in complex engineering problems.