Hi, I'm Tejeswar Pokuri, I leverage cutting-edge technology to solve complex real-world challenges. Leading the AI Research team at RUGVED Systems, I drive innovative research in Generative AI and autonomous systems, with a portfolio spanning groundbreaking projects like satellite image dehazing and assistive navigation technologies for the visually impaired. My professional journey includes impactful internships at TensorGo Technologies, where I optimized deep learning models, and notable recognition at premier conferences like NeurIPS and ACCV. Proven through competitive successes including the OpenCV AI Competition 2023 and finalist positions in JP Morgan and Goldman Sachs hackathons 2024, I bring comprehensive expertise across programming languages (Python, Java, C, C++), AI frameworks (TensorFlow, PyTorch, OpenCV), and web technologies (HTML, CSS, JavaScript, React, MongoDB), consistently pushing the boundaries of technological innovation and practical AI applications.

Working with Portfolio Performance Analysis team using Spring Boot, R, Oracle SQL, Grafana, Spark.
.jpg)
Leading a team of 15 skilled AI researchers, I am driving innovative projects focused on advancing research for prestigious conferences and developing an autonomous all-terrain bot. Our work has achieved significant recognition, with papers accepted at NeurIPS 2024 and ACCV 2024, reflecting our commitment to excellence in AI research and development.

Contributed to crowd counting, footfall analysis, lip syncing, video news analytics, and data analysis projects, achieving an impressive 85% accuracy in news video analysis through advanced deep learning and Quantized Computer Vision Models, reducing CPU inference time by 70%, and worked extensively on YOLO, Whisper, Wav2Lip, P2PNet, and innovative Video Moment Retrieval models.
.jpg)
I've worked on human anomaly detection in outdoor settings and developed an auto-navigation system for the visually impaired. Led a team of four at the OpenCV AI Competition 2023. Mentored 10 juniors in machine learning, exploratory data analysis, OpenCV, and Python. Conducted research on AI applications in the aerospace industry and optimization algorithms. Currently focused on Spacecraft Pose Estimation and Dehazing images.
.jpg)
.jpg)
Designed a comprehensive Navigation System catering to the visually impaired, integrating modules for Obstacle Detection, Depth Estimation, Scene Recognition, Barrier Detection, Facial Recognition, and Navigation. Employed customized YOLOv7 with a mean average precision (mAP) of 0.76 for household, road, and stairs detection, MIDAS for depth estimation, VGGFace for facial recognition, and EffNetB2 for scene recognition (12 classes, F1 score of 0.91). Leveraged MapQuest API for directions and achieved real-time navigation through the utilization of Nvidia GeForce GTX 1650 GPU.
View Project
Improved an open-sourced Hotel Automation Software featuring a user-friendly GUI, integrated chatbot, security surveillance system, and advanced data analytics, powered by PHP and MySQL.
View Project
Designed and implemented a movie booking system using a user-friendly Tkinter GUI and SQLite3 as the database. Developed a custom database structure adhering to 3NF standards and incorporated triggers for enhanced functionality.
View Project
Developed an Arduino Uno-based system using sensors and logistic regression to predict plant watering needs. Prototyped the solution in TinkerCAD, demonstrating real-time monitoring and automated decision-making.
View Project
Engaged in the development of Project Garuda, a smart surveillance system dedicated to human anomaly detection in outdoor regions. Implemented Mediapipe for anomaly detection, specifically targeting activities such as Running, Crawling, and Jumping.
View Project
Our team placed Top 7 in the OpenCV AI competition 2023, contributing to the development of Guiding Gaze—a cutting-edge navigation system for the visually impaired. The project received a $1000 prize, highlighting our team's commitment to excellence. Proudly represented India as the sole Indian team in the prestigious Top 7, showcasing global expertise and ingenuity.
More Details
Secured National Finalist position among 15,000+ applicants in JP Morgan's prestigious Code for Good 2024. Collaborated on solving the problem statement for U&I Trust, delivering impactful and innovative solutions.
Achieved National Finalist status among 3,000+ participants in the Goldman Sachs India Hackathon 2024. Developed innovative solutions for quant-focused problem statements using advanced data structures, algorithms, statistical modeling, and deep learning techniques.
Selected among 12,000+ applicants in a competitive online round featuring DSA and ML-based MCQs. As part of the prestigious summer school, gained in-depth knowledge in Reinforcement Learning (RL), Generative Adversarial Networks (GANs), and eXplainable AI (XAI).
Achieved second place in poster and paper presentation category at iACT 2023, hosted by International Association of Automation Bangalore section.
More Details
secured first place at the Intellect Connect competition, where I presented a paper titled 'GESSURE: A Robust Face-Authentic Enabled Dynamic Gesture Recognition GUI Application.' The competition focused on recognizing outstanding presentations, evaluating both content and delivery.
More Details
Secured first place in an ML hackathon by tackling an imbalanced dataset challenge with an innovative and efficient pipeline.
More Details
Developed AnimalHaze3k and IncepDehazeGAN to enhance wildlife imagery under haze, boosting detection accuracy by over 100% and advancing computer vision for conservation.
Developed a hierarchical distillation and augmentation strategy that boosts animal track classification accuracy by up to 12.4%, setting a new benchmark on the OpenAnimalTracks dataset.
Our paper is focused on using Self Supervised Learning approach for Dental Radiographs Segmentation using only 20% dataset and achieving better accuracy than many models.
Our paper is focused on using eXplainable AI approachs for epilepsy detection from EEG Signals.
Our paper is focused on Dehazing Satellite Images using a custom deep learning architecture and achieved highest accuracy on 2 publicly available dataset.
Our paper is focused on Detailed Systematic review on applications of AI in Aerospace.