Chat system
Summary
Built a Chat system that allows you to chat directly with your MySQL Database using natural language (Stack includes Langchain, Python, LLMs)
About
Work
Gradehive AI
|Founding Software Engineer
Summary
GradeHive AI leverages advanced reasoning Large Language Models (LLMs) optimized specifically for coding and grading tasks. Our AI helps streamline assignment creation, evaluation, and feedback, saving educators valuable time where users can create their own criteria/rubrics for grading
Highlights
Led the complete development of the frontend and backend systems for the product along with integrating LLMs wherever needed. Stack - ReactJS, NextJS, LLM, OpenAI, MongoDB, GitLab, AWS.
Features include automatic bulk assignment creation and auto grading using AI using a rubric, supporting 10+ coding languages (.py, .cpp, .js, .java, etc.), Handwritten JPEGs, Docs & PDFs using OCR.
Our feature set includes grading textual/imagery tests (or) code/computer science assignment based tests (or) data science/jupyter notebook based tests, where the AI not only grades but also speaks about why it's graded that way and which piece of the answer adds up to that grade.
Tooljet, Open Source
|Software Engineer Intern
Summary
Platform for Internal Tools
Highlights
Led the complete revamp for Tooljet's workflows feature, which is the backbone of the Tooljet platform, and also managed the transition of this feature from beta to production, while swiftly resolving production bugs/critical issues and ensuring seamless functionality. This upgrade was pivotal in securing two major enterprise customers, highlighting its critical role in our product's success.
Contributed to both frontend(ReactJS, Redux,Typescript,ReactFlow, SCSS ) and backend( Python, FastAPI)
Actively responded to customer feedback by creating Proof of Concepts for feature requests and pushing them to production, demonstrating our commitment to innovation and customer satisfaction, and picked up Open Source issues and fixed them.
Hoppscotch, Open Source
|Developer Intern
Summary
Alternative to Postman
Highlights
Revamped the whole marketing website - used React.js and Vue.js, and also Python for Backend,
Represented the organization's DevRel expertise at conferences and events, solidifying the position as leaders in API Testing.
Used SQL & Python libraries like Numpy, Pandas, and Matlplotlib to perform exploratory data analysis and find insights on developer testing.
Awards
Pupil on Codeforces (1220)
Awarded By
Codeforces
Pupil on Codeforces (1220)
4th Position in Opin Hacks(MLH) - India Wide Hackathon
Awarded By
Opin Hacks(MLH)
Secured 4th Position in Opin Hacks(MLH) - India Wide Hackathon
3rd Position in GDSC WOW - India Wide
Awarded By
GDSC WOW
Secured 3rd Position in GDSC WOW - India Wide
Finished 5 levels of Google Foobar
Awarded By
Google Foobar
Finished 5 levels of Google Foobar
1st Position in Worthy Hack- India-Wide Hackathon
Awarded By
Worthy Hack
Secured 1st Position in Worthy Hack- India-Wide Hackathon
3 Star on Codechef (1695)
Awarded By
Codechef
3 Star on Codechef (1695)
Knight on Leetcode (1871)
Awarded By
Leetcode
Knight on Leetcode (1871)
Speaker - Google DevFest 2024, India FOSS, TestMu Conference.
Awarded By
Google DevFest, India FOSS, TestMu Conference.
Speaker - Google DevFest 2024, India FOSS, TestMu Conference.
Global Rank 1402 in Round A Google Code Jam 2023
Awarded By
Google Code Jam
Global Rank 1402 in Round A Google Code Jam 2023
Skills
Languages
Python, Javascript, Typescript, SQL.
AI
LLM APIs, Langchain, MCP ( Model Context Protocol ).
Frameworks
ReactJS, Redux, VueJS, NextJS, NodeJS, ExpressJS, FastAPI, Flask.
Databases
Firebase, Supabase, MongoDB, RDBMS ( MySQL, PostgreSQL, SQLite).
Miscellaneous
Data Structures & Algorithms, REST, Linux, Git & GitHub, Kafka( Basic ), OpenAI API, Gemini API, Open Source, Machine Learning & Deep Learning Frameworks, Exploratory Data Analysis..
Concepts
Computer Networks, Object Oriented Programming, Operating Systems, DBMS.
Projects
Integrations b/w Claude Desktop and Brave Search & also Claude Desktop and TwentyCRM
Summary
Built integrations b/w Claude Desktop and Brave Search & also Claude Desktop and TwentyCRM to directly interact with these systems using Claude as a client using MCP ( Model Context Protocol ) (Stack includes Python & MCP )
Tooljet Workflows ( Internal)
Summary
Tooljet Workflows ( Internal) : Automate complicated manual business processes with less engineering effort. Achieve better business outcomes with the enterprise grade workflow builder( https://www.tooljet.com/workflows) - ReactJS, ReactFlow, ExpressJS, SCSS,Redux, TypeScript.
Spiking Neural Networks on the DVS Gesture 128 Dataset
Summary
Spiking Neural Networks on the DVS Gesture 128 Dataset : Spiking Neural Networks (SNNs) utilize discrete spikes to emulate biological neural communication. This study implements an SNN trained on the DVS Gesture 128 dataset using the snn Torch library. The network employs a single-layer architecture with Leaky Integrate-and-Fire (LIF) neurons, achieving 87.50% accuracy.