
Vishal Agarwal
Development
Karnataka, India
Skills
Machine Learning (ML)
About
VISHAL AGARWAL's skills align with IT R&D Professionals (Information and Communication Technology). VISHAL AGARWAL has 6 years of work experience, with 4 years of management experience, including a low-level position.
View more
Work Experience
Freelance Consultant
LLM Chatbot
May 2023 - Present
- An LLM RAG-based at an enterprise scale from the ground up. System capable of accommodating and managing additional billions of files maintaining low hallucination and high precision Competitor Data Analytics: In-depth data analytics on a well-funded startup's competitors, aiding in the formulation of a robust expansion strategy
Research Scientist - Lead Data Science
Rakuten Inc
December 2018 - April 2023
- Multi-task Learning using Vision Transformers & Knowledge Distillation (Res.): Conducted research on multitask learning using vision transformers and knowledge distillation; Filed 2 patents & presented research paper accepted in WACV 2023. Image Based Product Matching: Trained and deployed a SOTA Image Similarity Model based on Vision Transformers using Big DATA, enabling seamless processing of 3M images / week on a single machine. Proteomics Analysis: Oversaw the design of an ML-based predictive model for evaluating patient responses in cancer drug trials, identifying crucial biomarkers that distinguish responding from non-responding patients. Credit Card Fraud Detection: Led the training and development of a Deep Neural Network Architecture for credit fraud detection, achieving a 60% improvement in fraudulent transaction detection compared to traditional rule-based models. Additionally, managed the analysis of a large dataset, optimizing it to 1/100th of original size.
Research Intern - Data Science
Rakuten Inc
July 2018 - December 2018
- AI Diagnostics: Tele-medicine: Conceptualised & developed a cutting-edge deep learning-based diagnostics platform using image data from radiology, pathology, and dermatology. Successfully opened a collaborative medical data sharing channel with a renowned Multi-Speciality Hospital.
Summer Code Developer, Inkscape
May 2017 - August 2017
- CSS3 Parser: Enhancement of a CSS vector editor parser, elevating it from CSS2 compatibility to full CSS3 compliance.