SIDDARTH
NARAYANAN
ABOUT ME
Hey, I’m Siddarth Narayanan! I’m a student at Rutgers University pursuing a dual major in Computer Science and Data Science and a minor in Physics. I love challenging myself both mentally and physically, so you'd often find me playing sports like volleyball or basketball in my free time. I love learning about new topics and perspectives, so feel free to hit me up on LinkedIn!
PROJECTS
DIGITLEARN
I developed a continuous learning system that allows a user to draw digits using PyGame and train a neural network built with TensorFlow on their handwriting.
GITHUB
PYDIFFIE
A client-server implementation of the Diffie-Hellman key exchange algorithm built using sockets and secrets in Python.
GITHUB
SIDDARTHNARAYANAN.ME
You're looking right at it! This website was built using SvelteKit as my frontend, NodeJS as my backend, Sanity as my content management system, GSAP for those buttery smooth animations, and Vercel for hosting! The ultimate portfolio tech stack!
UNCLOCKED
One of my first projects, Unclocked was a task management app written in Flutter with Firebase NoSQL as the backend. This project taught me the complexities of state management and proper software design methodologies like separation of concerns.
01
DATA SCIENTIST INTERN
May 2024 - Aug 2024
During my time at Otsuka I leveraged GenAI to solve two main projects. First, I created an end-to-end GenAI application for document validation. I used technologies like Pydantic, MongoDB, FastAPI, ChromaDB, and Ollama. I also finetuned an LLM to rate manager feedback. I employed Parameter-Efficient Fine-Tuning (PEFT) techniques like LoRA and QLoRA, to efficiently train the model on Otsuka personal data. Feedbacks were desensitized using a custom Named Entity Recognition model built with SpaCy before being fed into the model.
GENAI
PYTHON
PYTORCH
COLLABORATION
NETWORKING
02
UNDERGRADUATE RESEARCH
Oct 2023 - May 2024
I researched how LLMs can be used to evolve code in the context of solving mathematical problems. A lot of mathematical problems are defined by concrete characteristics, but finding a solution that fits these characteristics can be challenging. My group and I used LLMs to improve functions to solve the Trifference Problem for large values (n > 9). Although we were unable to get results for n > 9, we saw a dramatic speed improvement for some values n < 9 compared to the brute force method.
PYTHON
HIGH COMPUTE
GENAI
CRITICAL THINKING
COLLABORATION
03
HEAD OF DESIGN, TONDOVA
Aug 2023 - PRESENT
Relational note taking is human note taking. Tondova aims to represent data as relational nodes in a large graph of knowledge. We aim to bring the power of AI to truly companion the human mind in creation, synthesis, and storage of information.
FIGMA
UI/UX DESIGN
LEADERSHIP
FRONTEND DEVELOPMENT
04
FREELANCE PROGRAMMER
Mar 2022 - Jun 2022
I developed an automated bot that would execute online bookings for a client in milliseconds, outperforming all competing registrants. I deployed the application to the Google Cloud Platform and used Cloud Scheduler + Serverless Cloud Functions to run the application weekly.
PYTHON
APIS
GCP
ITERATION
NETWORKING
05
DIRECTOR, CODIVATE
Jan 2020 - Aug 2023
As a director of development, I completely overhauled the intermediate python curriculum. I was also in charge of managing online zoom classes and interviewing new student tutors.
LEADERSHIP
PYTHON
COLLABORATION
06
VARSITY VOLLEYBALL CAPTAIN
Mar 2023 - Jun 2023
As the captain of my high school varsity volleyball team, I helped motivate my teammates and learned what it means to be an impactful leader. I realize that leadership is about understanding context and taking action based on common goals. It is also about remaining calm during stressful situations—something that I learned thoroughly as a captain.
LEADERSHIP
COMMUNICATION
PATIENCE
EXCELLENCE
TENSORFLOW DEVELOPER PROFESSIONAL CERTIFICATE
PCAP: PYTHON CERTIFIED ASSOCIATE
SEAL OF BILITERACY IN FRENCH
CERTIFICATIONS