Ramavath Suresh
ML EngineerAI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI & Autonomous Driving Engineer @ RoshAI || ADAS || Multi Sensor Fusion || 3D Perception || SLAM || Tracking || Planning || VLMs || LLMs || ROS1/ROS2 || CARLA .|IIT KGP’24
I’m an AI &Autonomous Driving Engineer at RoshAI, with an Integrated M.Sc. in Physics from IIT Kharagpur. I specialize in perception, sensor fusion, and map-based planning across the autonomous driving stack. My work focuses on multi-sensor perception using LiDAR, radar, and cameras, HD map integration, and localization (GPS/IMU, SLAM). I’ve developed end-to-end planning and auto-parking systems, enhancing spatial intelligence for robust autonomy. Proficient in Python, C++, ROS 1, ROS 2, TensorFlow, PyTorch, and CARLA, I’m passionate about creating intelligent, real-time mapping and perception solutions that power the future of autonomy.
Indian Institute of Technology, Kharagpur
Final year undergraduate student, Integrated physics
January 1, 2019 – April 1, 2024
RoshAi
Robotics and AI solution Engineer
June 1, 2024 – Present
ProfessorAI
AI Developer
March 1, 2024 – May 1, 2024
Remote
Asilla
Computer Vision Research Engineer
June 1, 2023 – August 1, 2023
Tokyo, Japan · Remote
Verzeo
Artificial Intelligence Intern
June 1, 2021 – August 1, 2021
Hyderabad, Telangana, India
Chegg Inc.
Chegg Subject Expert
June 1, 2020 – March 1, 2024
Remote
Resolving Missing Frames in Human Trajectories by Developing Reconstruction and Imputation LSTM Model
June 1, 2023 – August 1, 2023
•Developed a LSTM-based architecture meticulously designed to proficiently handle missing frames in complex human trajectories •Incorporated attention mechanisms, substantially elevating the effectiveness of encoding, decoding, and imputation processes •Achieved highly accurate trajectory prediction by leveraging the capabilities of a carefully constructed reconstruction decoder mechanism •Introduced a multi-head attention imputation decoder, effectively addressing intricate scenarios involving missing data patterns •Demonstrated the remarkable performance of the model, underscoring its excellence with an impressive R-squared value of 0.92
Fashion Recommender System with Deep Learning
June 1, 2022 – July 1, 2022
•Built an integrated fashion recommender system employing a fusion of deep learning architectures, and user interfaces •Utilized pre-trained ResNet-50 convolutional neural network for feature extraction, converting image features into embeddings •Engineered an interactive user interface (UI) using Streamlit for personalized fashion recommendations based on extracted features •Implemented a Nearest Neighbors algorithm for precise recommendation and efficient image similarity search
Theft Prevention using AI
June 1, 2021 – July 1, 2021
Built a theft prevention using AI and computer vision which will help an individual or a group of family members to get notified about the intrusion into the house/property. The system should recognise the family members (at least 1) as innocent and anyone apart from recognised faces will be treated as an intruder. Once an intruder is identified the system should trigger an email with the picture captured. •Created a model for house/property intrusion detection, focusing on prompt alerts for timely response to potential threats •Developed a Keras-based Convolutional Neural Network featuring a sequential of convolutional layers for enhanced performance •Incorporated a Haar cascade file into the model, boosting facial recognition accuracy and reinforcing security measures •Integrated Pygame to activate alarms when necessary, enhancing the system's ability to alert home owners effectively •Accomplished accuracy level of 0.88, utilizing categorical cross-entropy loss function to fine-tune the model's performance
Predicted users gender based on Twitter profile information using ANN
June 1, 2021 – June 1, 2021
•Engineered a model utilizing Artificial Neural Networks for precise gender classification from Twitter profile data •Analyzed data with Scikit-learn, Matplotlib, Seaborn, and Pandasto extract insights into the dataset's intricate features •Individually fine-tuned using MLP, RandomForest Classifier and SVM on TwitterData, making substantial contributions to prediction •Achieved an accuracy of 0.90 in the gender classification task, highlighting the remarkable efficacy of the engineered model
Cultural Fit Analysis
The candidate's projects are primarily academic/personal, focusing on core ML/AI problems. The experience at RoshAi and ProfessorAI aligns well with an ML Engineer role, indicating a practical application of skills. However, the overall experience is relatively short, and the projects, while technically sound, are mostly individual efforts. This suggests a potential fit for a technically driven culture, but further assessment would be needed for team collaboration and broader organizational fit. The focus on research and development in projects and internships aligns with an innovative, problem-solving culture.
Soft Skills & Operational Fit
The candidate's project descriptions indicate an ability to work on complex problems and deliver measurable results (e.g., R-squared 0.92, accuracy 0.90). The diverse project portfolio suggests adaptability and a proactive approach to learning and applying new technologies. However, without specific psychometric or English test scores, a detailed assessment of soft skills and operational fit is limited.