Computer Science Image Processing
Resume:
Sai Ram Kasanagottu
+1-631-***-**** ******.***********@**********.*** LinkedIn GitHub Google scholar Research Gate ORCID EDUCATION
SUNY Stony Brook University, New York 2024-Present Ph.D Student, Computer Science
Indian Institute of Technology, Kharagpur 2023
Master of Science (Research), Computer Science and Engineering CGPA 9.53/10 MS Thesis: Developing Reusable, Speedy, and Compact Deep CNN Models Indian Institute of Technology, Kharagpur 2014
Bachelor of Technology, Electronics and Electrical Communication Engineering CGPA 7.95/10 SKILLS
Programming: Python, C, C++, Java, Julia, HTML, CSS, Latex, Verilog, MATLAB Frameworks:VLM and LLM (OpenAI, Claude, Langchain, Llamaindex, ChromaDB, Mistral, LLama, ProteinBert), Huggingface, OpenCV, PyTorch, Tensor flow, Caffe, Matplotlib, Numpy, Pandas, ROS, Gazebo, FastAPI, OpenMP, MPI Environment:Visual Studio Code, Sublime, Text, MATLAB, Simulink, SVN, GIT Systems:Azure, AWS (Bedrock, Sagemaker, EKS etc), CentOS, MAC OS, Windows, Linux PUBLICATIONS
Conference Papers
A. Saha, K. S. Ram, J. Mukhopadhyay, P. P. Das, and A. Patra. Fitness based layer rank selection algorithm for accelerating CNNs by candecomp/parafac (CP) decompositions. In 2019 IEEE International Conference on Image Processing (ICIP), pages 3402–3406, Sep. 2019.
SaiRam, K., Jayanta Mukherjee, Amit Patra, and Partha Pratim Das. HSD-CNN: Hierarchically self-decomposing CNN architecture using class specific filter sensitivity analysis. In the Proceedings of the 11th Indian Conference on Computer Vision, Graphics and Image Processing, ICVGIP 2018, New York, NY, USA, 2018. Association for Computing Machinery. B. Saha, K. Sai Ram, J. Mukhopadhyay, A. Roy, and A. Navelkar. Video based person re-identification by re-ranking attentive temporal information in deep recurrent convolutional networks. In 2018 25th IEEE International Conference on Image Processing (ICIP), pages 1663–1667, Oct 2018.
J. Mukhopadhyay and K. Sairam. Nonseparable filters for images in the block DCT domain. In 2018 25th IEEE International Conference on Image Processing (ICIP), pages 1493–1497, Oct 2018. Patents
Wajahat Qadeer, Rehan Hameed, Satyanarayana Raju Uppalapati, Abhilash Bharath Ghanore, and Kasanagottu Sai Ram,
“Method for automatic hybrid quantization of deep artificial neural networks,” June 10, 2021. US Patent App. 17/112,889 RESEARCH EXPERIENCE
Project Officer, Indian Institute of Technology, Kharagpur October 2017– March 2019 Synopsys – IIT Kharagpur Collaborative Research, Supervisor: Dr.. Jayanta Mukhopadhyay, Dr. Partha Pratim Das, and Dr. Amit Patra
Project Title: Generic object detection and recognition techniques for embedded vision systems. Key roles:
● Designed and implemented a hierarchically self-decomposing CNN and Sub-networks as a fast, compact, and reusable deep learning model for low-power embedded applications.
● Designed and implemented a converging fitness-based layer rank selection algorithm for accelerating CNNs by the Tucker-2 CP decomposition method, keeping overall accuracy drops within limits.
● Successfully mentored two students and guided them toward publications at a flagship IEEE conference, the International Conference on Image Processing.
Junior Research Fellow, Indian Institute of Technology, Kharagpur December 2014 – September 2017 MHRD - CSE, IIT Kharagpur, Supervisor: Dr. Jayanta Mukhopadhyay and Dr. Partha Pratim Das Project Title: Autonomous Navigation of ground robots in aid of Drones. Key roles:
● Designed aerial vehicles to detect moving objects on ground in the first phase and extended it to follow and land on it.
● Implemented a state estimation algorithm based on optical flow to generate odometer data.
● Develop AI-based path planning in a Gazebo-based simulator to combat obstacles. Bachelor’s Student, Indian Institute of Technology, Kharagpur August 2013 – June 2013 Kalpana Chawla Space Technology Cell, Supervisor: Dr. Kalyan Bandyopadhyay BTech Thesis: Navigational Signal Processing of GNS signal to estimate user location. Key roles:
● Implemented Verilog in extracting ephemeral data to determine the user’s location.
● Investigated and implemented error correction models to minimise location uncertainty and optimise processing time. WORK EXPERIENCE
Saama Technologies Inc, Pune, India Aug 2023 – Aug 2024 Designation: Technical Lead Engineer
Key roles:
● Developed ETL pipeline (with data error handling capabilities) to extract Key topics from clinical protocol and SAP documents employed with various NLP techniques for Generative AI application model endpoint for generating headers from Clinical Data documents..
● Designed a system using an LLM model to effectively use the key information extracted from these documents to map with a reference global template to effortlessly generate TLF data with No-code and Drag and drop simplicity designs.
● Designed prototypes to efficiently streamline difficult process in clinical data management and approval process like automated SDTM generation by achieving 50% effort reduction and nearly 69% faster timelines,
● Worked on various prompts from sequential, Chain of thought, step back prompts in an instruction set for extraction in creating design for automated siteline presentations, Site marketing mails, TextQueryBots with SQL, Python and R integration capabilities.
● Fine-tuned Llama2 and Mistral LLM models to the clinical domain.
● Worked on amazon bedrock Gen AI services (Claude AI LLM)and Azure OpenAI API (GPT3.5 models) for extracting structural information and also for generation using Langchain and LLamaIndex frameworks. DisplaySweet, Melbourne, Australia March 2021– November 2022 Designation: Lead ML/ Deep Learning/ Computer Vision Engineer Key roles:
● Serve high-end applications in the digital real estate industry, numerous deep state learning models and computer vision algorithms including hotspot generation, semantic floor plan and level plan region-icon segmentation, optimization of 3D building mesh, and information extraction from building plan photos using OCR
● Processing and cleaning of the internal data to maximise purchasers’ interest in land properties.
● Integrated project schedule and duration estimations, which increased business productivity and efficiency.
● Developing an AI agent to help internal members with property recommendations and queries for building customer segment analysis, query processing in search.
Kinara, Inc, Los Altos, CA - MLAI System, Hyderabad November 2018 – August 2019 Designation: Deep Learning Software Engineer
Key roles:
● Designed and incorporated an automated algorithm for mixed bit quantization and pruning of deep learning models into their existing end-to-end Compiler workflow, resulting in a patent filing.
● Resolved various issues related to the integration of video-based segmentation and pruning for edge computing. Philips Electronics India, Noida, India July 2014 – October 2014 Designation: Assistant Manager - LED Development
Key roles:
● Prototyped design of a scene switch LED bulb, a native IoT-based low-segment device solution.
● Collaborated with industry manufacturers to mass produce prototypes and perform extensive testing and modifications to qualify them to the industry standard.
RESEARCH INTERNSHIPS
Research Intern, Supervisor: Dr. Prabir Kumar Biswas, Indian Institute of Technology, Kharagpur May 2013– June 2013 Project: Effective Panoramic image generation from video sequence. Research Intern, Supervisor: Dr. Subhashish Bhasak Variable Electron Cyclotron Centre, Kolkata June 2012 – July 2012 Project: Design and Implementation of IQ phase shifter for LLRF control RELEVANT COURSEWORK
Introduction to Computer Vision, Machine Learning, Advanced Digital Image Processing and Computer Vision, Pattern Recognition and Image Understanding, Digital Signal Processing, Algorithm Design and Analysis, Robotics, Satellite Communications Systems, Digital Voice and Picture Communication, Probability and Stochastic Processes, Matrix Algebra.
TEACHING ASSISTANTSHIP
1.Foundations of Computer Science, Stony Brook University Fall 2024 2.Advanced Image Processing and Computer Vision, IIT Kharagpur Spring 2018 3.Digital Image Processing, IIT Kharagpur Fall 2015, Spring 2017 4.Computer Graphics and Laboratory, IIT Kharagpur Spring 2016 5.Programming and Data Structures, IIT Kharagpur Fall 2016 FELLOWSHIPS & AWARDS
1.Awarded with Stony Brook University Computer Science Chairman Fellowship Award, 2024-25. 2.Junior Research Fellowship of Ministry of Human Resource Department (MHRD), Government of India as a research scholar in Indian Institute of Technology Kharagpur of Rs. 1 million. (2015-2017) 3.Best Team cooperation award at International Aerial Robotics Competition 2016 Beijing, the longest running autonomous competition in the world along with an Institute grant of Rs. 65K. (2016) 4.Recipient of Merit cum Means Scholarship (Rs. 60K/year) awarded by Indian Institute of Technology, Kharagpur for consecutive three years (2011 – 2013)
5.Won 4th position in Hardware Modelling at Inter Hall event, Technology General Championship, IIT Kharagpur. (2013) 6.Finals qualifier at ANADIGIX, an electronic circuit design challenge in Kshitij, techno festival at IIT Kharagpur. (2013) 7.Top 1% in National Standard Examination in Physics, conducted by Indian Association of Physics Teachers. (2010) 8.Finals in Indian National Chemistry Olympiad Exam, conducted by Homi Bhabha Centre for Science Education. (2010) LEADERSHIP EXPERIENCE
1.Senior Software Coordinator, Ariel Robotics Kharagpur 2015-2017 2.Captain, Software team of Hardware Modeling, Lala Lajpat Rai Hall of Residence, IIT Kharagpur 2013-2014 3.Treasurer, Board of Director, GYWS, an NGO, IIT Kharagpur 2013-2014 4.Executive Member, GYWS, an NGO, IIT Kharagpur 2010-2012 REFERENCE
Prof. Jayanta Mukhopadhyay
Professor, Computer Science &
Engineering, IIT Kharagpur
***@***.******.**.**
Prof. Partha Pratim Das
Professor, Computer Science &
Engineering, IIT Kharagpur
***@***.******.**.**
Dr. Shashaank Mattur Aswatha
Assistant Professor, Electrical,
Electronics, and Communication
Engineering, IIT Dharwad
********@*****.**.**
Contact this candidate