Hello!

I am Ankitha Suresh

Student, Developer & Wanderlust

About Me

Who am i?

Hey there! 👋 I'm a Master's student in Computer Science at UMass Amherst, and I've worked as a software engineer at HPE. I love diving into code, building algorithms, and creating cool tech solutions.

I'm passionate about building excellent software that improves the lives of those around me.

When I'm not studying or coding, you'll probably find me checking out the latest tech trends, solving puzzles, traveling, or watching a good thriller movie.

My Skills

What am I good at?

Python
90%
Java
80%
HTML & CSS
85%
Javascript
90%
REST
85%
System Design
90%

Experience

Headstarter 07/2024 - 08/2024

Software Engineer Fellow

  • Designed CRM integration API endpoints for Microsoft Dynamics 365, automating bidirectional sync of data across 6 modules that processed 50000+ records monthly, reducing manual updates by 85%.
  • Built 3+ full-stack applications using React, integrating large language models (LLMs) for AI-driven features, enhancing user engagement and application performance.
  • Innovated a responsive pantry management system using HTML, CSS, JavaScript, AWS CDK, and MySQL, enabling efficient tracking of 1,000+ inventory items, and incorporated predictive analytics for real-time stock predictions, attained 500 Mbps throughput via load balancing.

Hewlett Packard Enterprise 09/2021 - 07/2023

Software Engineer

  • Developed a RESTAPI in Golang for high-availability automation for Oracle, SQL Server, and SAP, reducing latency by 45% at 10K RPS while maintaining 99.9% uptime SLA.
  • Achieved 300% faster cluster failover (8s to 2s) by revamping the I/O scheduler in C with lock-free queues, enhancing high-availability across 50+ production nodes.
  • Accomplished successful development of a high-availability cluster management API, facilitating seamless automation of node failover. Integrated RESTful APIs with Serviceguard Manager, reducing manual intervention by 70%, supporting 50+ nodes in production.

Hewlett Packard Enterprise 02/2021 - 08/2021

Research and Development Engineer

  • Delivered an automated installation framework using Python, streamlining software deployment, and reducing setup time by 90%.
  • Prototyped a delta-encoded storage replication algorithm for high-availability clusters, reducing cross-site WAN bandwidth usage by 35%.
  • Devised Bash test suite for system validation, improving defect detection by 60% and catching 120 configuration errors pre-deployment.
My Work

what have I done?

Buddy: AI Chatbot Github

It is an intelligent conversational AI built with Next.js and React, using OpenAI’s models for accurate responses. It features Retrieval-Augmented Generation (RAG) to pull from a knowledge base, user login for personalized interactions, and a feedback mechanism for rating responses. Leveraging Pinecone for knowledge management and MongoDB for data storage, the application is efficiently deployed on Vercel.

Car Make and Model Classification Github

Engineered a highly efficient machine learning model using Convolutional Neural Networks (CNNs) to analyze a vast dataset of car images, achieving precise prediction of car make and model. By implementing the Inception-v3 model, I attained a 92% accuracy rate, which facilitated the development of an advanced image recognition system for automotive applications.

Portfolio GitHub

Developed a dynamic and interactive portfolio website using HTML, CSS, Javascript and PHP. This project showcases my skills and projects, providing an engaging user experience with smooth navigation and responsive design. The backend handles data efficiently, ensuring seamless content management and display.

Algocards GitHub

Algocards is an AI-powered tool that generates flashcards for mastering Data Structures and Algorithms (DSA). It offers three sets: Basic DSA, Algorithms, and Advanced DSA, along with assessments to test user knowledge. Built with Next.js and React, Algocards provides an interactive learning experience for DSA enthusiasts.

Crop Yield Prediction GitHub

Used Python, scikit-learn, and pandas to analyze a 5-year agricultural dataset. By adapting the Gradient Boosting Algorithm and employing feature scaling and selection techniques, I achieved a 91.8% accuracy in predicting crop yields. Additionally, I developed a Heroku-powered mobile app that determines suitable crops and predicts yield percentages based on user-provided location data, optimizing agricultural productivity by up to 75%.

Buddy: AI Chatbot Github

It is an intelligent conversational AI built with Next.js and React, using OpenAI’s models for accurate responses. It features Retrieval-Augmented Generation (RAG) to pull from a knowledge base, user login for personalized interactions, and a feedback mechanism for rating responses. Leveraging Pinecone for knowledge management and MongoDB for data storage, the application is efficiently deployed on Vercel.

Car Make and Model Classification Github

Engineered a highly efficient machine learning model using Convolutional Neural Networks (CNNs) to analyze a vast dataset of car images, achieving precise prediction of car make and model. By implementing the Inception-v3 model, I attained a 92% accuracy rate, which facilitated the development of an advanced image recognition system for automotive applications.

Portfolio GitHub

Developed a dynamic and interactive portfolio website using HTML, CSS, Javascript and PHP. This project showcases my skills and projects, providing an engaging user experience with smooth navigation and responsive design. The backend handles data efficiently, ensuring seamless content management and display.

Algocards GitHub

Algocards is an AI-powered tool that generates flashcards for mastering Data Structures and Algorithms (DSA). It offers three sets: Basic DSA, Algorithms, and Advanced DSA, along with assessments to test user knowledge. Built with Next.js and React, Algocards provides an interactive learning experience for DSA enthusiasts.

Crop Yield Prediction GitHub

Used Python, scikit-learn, and pandas to analyze a 5-year agricultural dataset. By adapting the Gradient Boosting Algorithm and employing feature scaling and selection techniques, I achieved a 91.8% accuracy in predicting crop yields. Additionally, I developed a Heroku-powered mobile app that determines suitable crops and predicts yield percentages based on user-provided location data, optimizing agricultural productivity by up to 75%.

Buddy: AI Chatbot Github

It is an intelligent conversational AI built with Next.js and React, using OpenAI’s models for accurate responses. It features Retrieval-Augmented Generation (RAG) to pull from a knowledge base, user login for personalized interactions, and a feedback mechanism for rating responses. Leveraging Pinecone for knowledge management and MongoDB for data storage, the application is efficiently deployed on Vercel.

Contact

Get In Touch!

413 472 0926
ankithasures@umass.edu