Hello!
I'm Atharva Patil, MSCS student at UMass Amherst specializing in building high-performance systems and AI-powered solutions to solve complex real-world problems.
Feel free to connect with me on my Linkedin or drop a mail on Gmail
My journey in technology began with a curiosity for how things work and a drive to improve them. Over time, that curiosity evolved into a passion for designing efficient systems and leveraging AI to create a meaningful impact.
At Ernst & Young, I leveraged advanced technologies to support clients in their ESG and Sustainability strategic planning, helping them track progress and achieve measurable goals toward long-term impact.
Beyond professional work, I've worked on personal projects ranging from a real-time collaborative application to a trading system simulator, each reinforcing the value of continuous learning and rapid problem-solving, while giving me hands-on experience with modern software technologies, spanning cloud platforms (AWS, Azure), frameworks (React, Angular, Django, Express), and tools such as Docker, Redis, and Neo4j."
Now, at UMass Amherst, I'm deepening my expertise in Machine Learning, generative AI and Natural Language Processing, while seeking opportunities to contribute to innovative teams building the next generation of technology.
- Python
- C & C++
- Java
- JavaScript
- TypeScript
- SQL
- PHP
- HTML
- CSS
- Express.js
- Django Rest Framework
- Django
- React
- Angular
- Flutter
- Next.js
- Docker
- AWS
- Microsoft Azure
- gRPC
- MongoDB
- Neo4j
- Redis
- Celery
- Firebase
- Postman
- Git
- Material UI
- Bootstrap
- Tailwind CSS
- Angular Material
At EY, I worked on enterprise-scale ESG (Environmental, Social, and Governance) reporting application used by some of India's largest corporate groups. My focus was on building scalable modules that improved data transparency, reduced manual effort, and strengthened security across the platform
- Value Chain Assessment Module: Led the end-to-end development of a module used by 4 enterprise clients to evaluate ESG performance for 400+ partner organizations. Built customizable survey forms with advanced scoring logic, enabling clients to assess sustainability KPIs with precision
- Role-Based Authentication: Integrated with MSAL and MyEY (EY's global IAM portal) to build a secure, role-based authentication system. Automated partner onboarding workflows, reducing manual effort by 80%
- Data Management Module: Designed a submission tracker and approval workflow to monitor ESG data completeness across all subject areas. This improved visibility and accountability for multiple stakeholders.
- Error Reduction & Validation: Implemented audit trails, validation rules, and a historical variance checker that reduced submission errors by 30% and flagged inconsistencies early.
- Efficiency Gains: Developed dynamic forms that automatically lists relevant ESG data fields by subject area, eliminating manual dropdown selections and cutting data entry time by over 60%
- Client Impact: Released digital ESG reports for 3 major clients in the energy and automotive sectors, improving transparency and compliance readiness.
- Digital ESG Reporting: Developed an interactive web-based ESG report with data visualizations for one of India's largest companies, delivering one of the first market-ready solutions to highlight sustainability advancements done by the company
- Cloud Deployment: Set up a CI/CD pipeline in Azure DevOps and deployed via Azure App Service, reducing deployment time and ensuring secure, seamless updates.
- AI Sustainability Chatbot: Built an AI assistant that helps clients and stakeholders find answers to sustainability-related queries with the use of GPT 3 API. Improved factual accuracy by fine-tuning the BART-LFQA (Bidirectional and Auto-Regressive Transformer for Long-Form Question Answering) model with data extracted from 40+ sustainability websites.
- Semantic Search Engine: Fine-tuned the RoBERTa (Robustly Optimized BERT Pretraining Approach) model on ESG reports to enable advanced semantic and contextual search, helping stakeholders retrieve relevant insights faster.
- Developed a streaming application with secure user authentication, ensuring data privacy.
- Designed and implemented a rating & review system to boost user engagement and collect actionable feedback.
- Built an admin portal to manage content release schedules, user accounts, and subscription plans, streamlining content operations and reducing manual effort.
- Developed a real-time video processing system to monitor pedestrian flow on a bridge using live CCTV feeds.
- Integrated IoT-based door control, publishing lock/unlock commands from the vision system to a subscriber service managing access control.
Ranked 2nd among 1000+ participants for developing a vision analytics solution that tracks global green cover changes over time using satellite imagery from the Google Earth Engine (GEE) dataset.
Secured a spot in the top 100 out of 1,000+ teams for building Monito, a responsive chatbot web app that monitors application health, security, performance, and potential issues.
Selected from 20,000+ entries across India for developing QBaG, a crowdsourced Question Bank Generator that enables academicians to generate objective questions and exam papers through an automated, robust system.
Built a collaborative platform that enabled real-time editing across text, drawing, and code, allowing multiple users to work together seamlessly . Designed an isolated and secure execution environment supporting multiple programming languages, while ensuring scalability and reliability through modular services and efficient inter-service communication. This enhances teamwork, streamlined knowledge sharing, and improved accessibility for developers and learners working remotely
Developed a vision analytics solution that tracks and visualizes changes in global green cover using aerial and satellite imagery. Users can explore vegetation density trends over time through NDVI and EVI layers, generate interactive time-lapse animations, and analyze the impact of vegetation changes on climate, and air quality. At a regional level, the system provided district-wise insights and predictive modeling of land cover types, helping policymakers and researchers better understand ecological shifts and plan sustainable interventions. This enables data-driven environmental monitoring and decision-making by transforming raw satellite imagery into actionable insights on deforestation, vegetation health, and climate patterns.
Developed a web-based engine that classifies, translates, and summarizes web pages from unknown URLs, extracting insights such as website's category (e.g. sports, entertainment, politics etc.), related topics, and similar websites. This significantly reduces the time needed to identify and understand the core information on unfamiliar websites.
Developed a crowdsourced platform for generating question papers from a large pool of objective questions. The system enables paper setters to create customized papers through an intuitive interface, selecting questions based on multiple parameters and exporting them in various formats. Each contributed question undergoes similarity check, automated validation and expert review, ensuring high-quality and reliable assessments. This reduces the effort and time required for paper creation while improving question quality, consistency, and scalability of exam preparation.
Built a high-performance Order Book system with support for 5 order types, multithreaded time-based order pruning, and O(1) order tracking. Ensured reliable trade execution and clean lifecycle management through robust data structures and comprehensive test coverage.
Built a chatbot that enables automated monitoring of application health, security, and performance. It allows users to track API endpoints and websites, receive instant alerts on potential issues, and generate detailed audit and security reports with visual insights. This reduces manual monitoring time and improves system reliability by providing automated alerts and actionable insights, helping users identify and resolve issues faster.