Utsav Doshi Email & Phone Number

New York, United States

New York, United States

Chennai, Tamil Nadu, India

• Developed an "AI-based Automated Descriptive Answer Evaluation System" aimed at reducing human intervention in answer paper evaluation by using BERT and BM25 models.• Designed to streamline and automate the process, the system sought to improve efficiency by using NLP models to assess descriptive answers. Created a mathematical model for testing validity, which was implemented as a parallel entity to traditional evaluations.• The system was introduced as a supplementary evaluation tool to enhance accuracy and scalability for evaluating 8,000 papers.

Delhi, India

• Focused on developing and maintaining websites with a particular emphasis on SEO/SEM strategies, SaaS tools for e-commerce, and AI tools tailored to client needs.• The goal was to enhance site performance, increase visibility, and improve user engagement for e-commerce clients. Employed SEO/SEM for higher organic traffic and developed customized AI tools to meet client requirements.• Achieved a 20% increase in traffic engagement and a 40% boost in customer sales productivity. Developed tools that doubled website traffic, reflecting significant improvements in client business metrics.

Redmond, Washington, United States

• Led a project on "Responsible Generative AI" under mentorship from Microsoft’s Sr. Director, utilizing the GLIDE model to generate original images without replicating existing artists' works.• Aimed at responsible AI use in creative fields, the project focused on creating ethically generated images. Weekly mentorship meetings provided insights into model refinement and ethical guidelines.• Developed a model capable of generating unique visuals, contributing to responsible AI practices in creative domains.

Mumbai, Maharashtra, India

• Deployed an AI-driven SD-WAN management system that dynamically adjusted configurations based on real-time network conditions to improve network performance.• Implemented to reduce downtime and enhance network efficiency, the AI system optimized WAN configurations based on live data. Additional tools analyzed traffic patterns for optimal data transmission.• Achieved a 25% reduction in downtime, a 30% improvement in network performance, 20% reduction in data latency, and significant cost savings in WAN operations (15% bandwidth and 10% operational cost reduction).

Singapore

• Collaborated on developing Bayesian model cross-validation techniques to enhance model accuracy.• The team identified key areas to boost accuracy by diagnosing and resolving data issues and used Bayesian methods for cross-validation.• Improved model accuracy to 97.73%, documented the findings in a research article that detailed methodologies and outcomes.

Masters In Computer Science, Computer Science

Relevant Coursework: CS-GY 6063 Software Engineering, CS-GY 6083 Principles of Database Systems, CS-GY 6513 Big Data

Bachelor Of Technology - Btech, Computer Science And Engineering With Artificial Intelligence And Machine Learning, 9.5/10 Cgpa

Activities and Societies: • Part of SRM's Directorate of Learning and Development, guided by the college director • Microsoft Learn.

Nursery -> Class 10Th -> Class 12Th