Vivek Gopal Ramaswamy Email & Phone Number

San Francisco, Ca, Us

San Francisco, California, Us

Develop Mask-RCNN based deep learning model for detection and quantifying of Amyloid Beta Plaques and Neurofibrillary Tangles to study the progression of Alzheimer's disease. Development of a foundational model for predicting the cell types (Neuronal vs Glial ) cells Implementation of gradient based clustering for effective data annotation using Active Learning Design and Development of Image processing pipeline using Nextflow and Dolphin Next Implement Ablation CAM technique based on Gradient Class Activation Map for explaining AI model predictions.Create a custom Qupath plugin for Annotation of Histopathology Slides and to perform inter-rater variability studies.Develop Graph Neural Network techniques for integration of spatial transcriptomics data with Image data.Implement unsupervised clustering algorithm using Meta’s state-of-art DinoV2 foundational Model to find new phenotypes from the Whole Slide Imaging data.Design and Development of AI based closed loop Microscope for high throughput experiments

San Francisco, California, Us

- Built a minimum viable product in 2 Months that helped dental providers collect patient payments 10X faster.- Developed and executed a successful go-to-market strategy for a new B2B payment platform, resulting in a 45% increase in the user adoption rate.Strategically devised and formulated new inbound marketing channels, yielding a substantial 17% increase in the website traffic.- Gathered customer feedback to inform development of automatic invoicing, resulting in a 28% increase in the revenue and 33% reduction in the average accounts receivable rate.- Acquired a $1.5M revenue-generating dental clinic as the first customer within the first week of launching.

Menlo Park, California, Us

- Led a 15-person cross-functional robotics team to build an AI based advanced lane overtaking system that increased the average speed of the vehicle by 20% and reduced autonomous vehicle disengagement by 60 percent.- As a Product owner, I Led a team of 10 Engineers to design and develop Vehicle to Vehicle (V2V) communication systems, which helped reduce dangerous maneuvers and increase customer safety by 70%.- Decreased software release time by 22% by implementing a rule-based auto-grading system to automatically verify Autonomous Driving Scenarios.

Pittsburgh, Pa, Us

45805-A Lean Entrepreneurship

Pittsburgh, Pa, Us

16-650 A Systems Engineering and Management for Robotics

Farmington Hills, Mi, Us

• Worked on 3D LiDAR annotation tool which could detect, track and annotate objects in 3D. • Implemented algorithm for conversion of Point Cloud Data to Birds Eye View (BEV) and vice versa.• Employed fusion of Images and LiDAR data in conjunction with a deep learning model for 3D object detection.• Prototyped a concept called “Deep Box”, using a feature extractor from VGG16 and ROI align layer from Faster R-CNN

Quadcopter for aerial inspection:• Developed an object recognition algorithm using Single Shot Multi-box detector(SSD) using CNN for inspecting important items in a factory.• Stabilized video from the quadcopter by implementing Kanade-Lucas-Tomasi(KLT) feature tracker algorithm.3D dense reconstruction of the Scene for Mapping • Worked and developed on a dense reconstruction model using the structure for motion(SFM) with epipolar and multiview geometry.

Riverside , California, Us

• Implemented object recognition algorithm using Single Shot Multi-box detector (SSD) for real-time vehicle detection.• Worked on Video Action Classification using ResNet18 and RNN with HMDB51 dataset.• Implemented key functions to analyze the model like visualizing conv layers, fc7 features, t-SNE projections.

Master'S Degree, Robotics Systems And Development (Mrsd)

Bachelor Of Engineering - Be, Electrical And Electronics Engineering

Education record