Ivan Nikitovic's repositories
stack-based-language
Typed stack language written in OCaml consisting of a compiler and interpreter, supporting scope, variable and function definitions and calls, conditional statements, mathematical operations, and recursion.
twitter-sentiment
This model implements the pre-trained TwHIN-BERT language model based on BERT, fine-tuned on the Sentiment140 dataset for the purposes of predicting sentiment of tweets on Twitter as either positive or negative.
cache-simulator
Cache simulator in C that supports load, store, and modify instructions, reporting hits and misses and handling evictions as projected
Count-Min-Sketch
Implementation of Count-Min Sketch using Pyhon. Count-Min Sketch (CMS) is a probabilistic data structure that serves as a frequency table of events in a stream of data.
CS506-Spring2023
Code and Slides for Boston University CS 506
job-tracker
Upcoming platform that helps candidates organize and manage job applications in tech. This application will provide users with tools to manage and monitor their job search process more efficiently than before.
seglist-memory-allocator
Simple and efficient memory allocator implemented using a segregated list. Implementation inspired by and partly taken from CS:APP by Randal Bryant.
sublinear
Python library implementing a subset of streaming algorithms. Includes variations of these algorithms (e.g. adversarially robust), as well as support for multiple data types.
user-interview-summarizer
Tool designed to help user researchers transcribe and summarize interviews easily. It leverages OpenAI's Whisper and GPT API to convert audio or video interviews into text and provide a summary of the content, as well as key observations and memorable quotes.
few-shot-chatgpt
Few-shot learning method using ChatGPT following the approach suggested in the paper 'A Neural Network Solves, Explains, and Generates University Math Problems by Program Synthesis and Few-Shot Learning at Human Level' https://arxiv.org/pdf/2112.15594.pdf
hub-and-spoke-mip
Optimization of a city delivery network using Gurobi for Python. Achieve efficient perishable goods distribution with a hub-and-spoke model, balancing delivery volumes and distances.
maximal-matching
Exploring the complexity of different versions of a graph exploration method for deciding whether an edge is included in a maximal matching.
text-to-graphics
We wish to demonstrate that a neural network trained on image-LaTeX pairs can effectively convert images from mathematical questions to input understandable by Codex using few-shot and zero-shot learning.