Jayant Kumar Email & Phone Number

California, United States

San Jose, Ca, Us

Multimodal Intent Understanding, Custom LLM tuning, LLM-based evaluations and Knowledge Graph based recommendations.

San Jose, Ca, Us

Responsible for user intent understanding modeling, asset and help search, discovery and recommendations, designing the technology roadmap and executing our vision for these novel features and experiences. I lead a team of ML engineers for ML pipeline/features in creative cloud products. My projects focus on creative content creation, semantic search/discover/recommendation, and assistive experience. We have delivered multiple ML based features in CC products including Creative Knowledge graph and its applications in recommendations/search.

San Jose, Ca, Us

I worked with multiple Sensei and Search Engineering teams.(1) ML/AI platform and building blocks to democratize and accelerate machine learning at Adobe. I lead big initiatives such as onboarding of large-scale ML datasets, containerized docker engines for training, evaluation, and batch processing.(2) ML for Search, Discovery, and Personalization.(3) NLP and Knowledge graph for Recommendations.

San Jose, Ca, Us

I played a crucial role in building the foundation of the Sensei ML framework. I conceptualized and developed the first custom trainable and dockerized engine in the platform with successful delivery of several use cases, which opened the door for basic AutoML capabilities in the framework. I made model development for many use-cases as easy as bringing relevant dataset to platform.I helped simplify the development workflow, interactive GPU-based development, dataset management, and other building blocks. I also contributed to projects related to fine-grained classification, face clustering, dynamic categories, and other model development.

San Francisco, California, Us

Helping aspiring machine learning engineers in capstone projects and finding ML/DS jobsI have mentored eight mentees so far on the following interesting topics:- Medical image segmentation (Computer Vision)- Text-based Question Answering (NLP)- Text Visual Question Answering (VQA)- Machine Translation (English -> Indian Languages)- Mobile-based sensor data analysis

Cupertino, California, Us

Individual contributor in the Deep learning-based perception team supporting autonomous technologies. Responsible for detection of two dynamic object categories. Field tested and presented our work on dynamic object detection to the SVP/Head of the project. Mentored interns on their summer projects.-Deep Neural Networks (DNNs) for 2D object detection and image segmentation (Tensorflow/Python). Deployed and tested the model in the field with an end-to-end pipeline.-Impact of augmenting training data with simulated images on DNN detection performance (Scripts to generate corner scenarios in the simulator and used it for training).-DNNs for range and 3D bounding box prediction to obtain pose for objects in RGB images (Tensorflow/Python).-Estimation of surface normals in the vicinity of navigating robot to infer yaw, pitch, and roll-Developed an evaluation library for metrics related to image segmentation and object detection (Python).-Dense CRFs for improving semantic image segmentation (C++).-Integrated/Deployed DNN models on the platform and performed field testing (C++).

Palo Alto, California, Us

Member of the Mobile Imaging team. -Deep neural networks for counting passengers in HOV/HOT lane vehicles (Project: Xerox Vehicle Passenger Detection System (VPDS)) (Caffe/Python). -Facial skin quality assessment and prediction (mobile app from P&G compass project)-PARC liaison for collaboration on egocentric vision and sensing with Georgia Tech. -Collaborator on university affairs committee project on computational photography with UMD. -Co-authored 17 patent applications and 3 conference papers

Palo Alto, California, Us

-Egocentric video analysis (hand segmentation, summarization, action classification, hand gesture recognition) and multi-modal data fusion-Video analysis features for Xerox Safe Courier app including flash/no-flash fusion, image quality check, video-based auto capture

- Developed novel methods for data selection of large data sets for SVM training [ICDAR11]- Developed an affinity propagation based method for text-line segmentation in handwritten document images [published at DAS10, ICDAR11]- Designed and developed efficient methods for No-reference Image Quality Assessment [CVPR12, CVPR13, CVPR14]- Proposed a multi-instance learning method for classification and localization of signatures and text in document images [ICDAR11]- Developed an application for structure based retrieval/classification of document images [ICPR12]

Norwalk, Connecticut, Us

-Developed an end-to-end prototype demonstrating the ability to extract high quality images from a mobile video capture of a multi-page document. -Conducted user study to evaluate user preference and difficulty level for such an application.-Developed and demonstrated a novel approach for separating overlapping handwritten and machine-printed text in document images.

Developed an efficient method for effectively estimating the sharpness/blurriness of document and scene images. The proposed method can be used to compute the sharpness in scenarios where images are blurred due to camera-motion (or hand-shake), de-focus, or inherent properties of the imaging system. I implemented the method in C++/Java on Android platform.

Arlington, Va, Us

I worked in the speech and language tech. lab. at Raytheon BBN technologies. Designed and implemented a shape codebook based approach for discriminating handwritten and printed text regions in Arabic document images (DRR 2011).

Bangalore, Karnataka, In

Project - Online Handwriting Recognition of Indian Languages(Funded by Ministry of Information Technology, India)I designed and implemented a tool using OpenCV for online handwritten data collection and automatic segmentation of handwritten data.Project : Field Extraction from Document Images(Funded by First Indian Corporation (FIC), Bangalore, India)I implemented modules in C for automatic field identification and extraction in document images.

Bangalore, Karnataka, In

Project : Infospace Mobile SearchImplemented change requirements for a product called Infospace Mobile Search ( C#, ASP.NET).

Houston, Texas, Us

Project - Parallelization of Samba on HP NonStop/VSSI proposed and implemented a scheduling algorithm for Samba, a file sharing software, so that the work load can be distributed across multiple processors.

Ms, Phd, Computer Vision

Bachelor Of Engineering - Be, Computer Science And Engineering