Kevin's Projects
ASR using CNNs and RNNs in PyTorch with the Connectionist Temporal Classification (CTC) algorithm
Autonomous Agent written using LangChain, OpenAI, and RAG
Code written while going through https://cdkworkshop.com/ - to build stacks, deploy an app on Lambda using DynamoDB, and write a CI/CD pipeline with CodePipeline
Solutions to the codesignal arcade problems
Conversational bot akin to ChatGPT made with LangChain, with OpenAI and HuggingFace Hub for LLMs
Lab assignments for Introduction to Data-Centric AI, MIT IAP 2024 š©š½āš»
Decoder-only Transformer model implemented from scratch in Python
Misc. Data Science projects I work on during my job hunt
Some basic data analysis on the data for electoral bonds released by the ECI
Solutions to problems from https://www.techinterviewhandbook.org/grind75 and other problems listed on the tech interview handbook. (Leetcode problems)
Config files for my GitHub profile.
Portfolio
Implementing basic stuff like LLMs and Prompting in LangChain
A simple langchain QnA pipeline trained on any local text using Sentence Transformers and OpenAI
Leetcode problems from the Grind75 set (https://www.techinterviewhandbook.org/grind75)
Llama 2 (7B) finetuned on 50k instruction-tuning data produced with GPT4
Llama 3 (8B) finetuned on Alpaca instruction-tuning dataset generated with GPT4
Course to get into Large Language Models (LLMs) with roadmaps and Colab notebooks.
A Python differ that returns diffs and their type for a pair of code snippets
Basic language model(s)
Small autograd engine
Miscellaneous data science projects and notebooks
ML and data analysis algorithm implementations (mostly) from scratch
Various music recommender algorithms
A small GPT inspired by nanoGPT from Andrej Karpathy
Small (<~1B) LMs for next token prediction in Python
Structured state space for sequence modeling (S4)