About Me
Hello, I am Aditya Jeswani, a first year Masters in Computer Science student at UC San Diego. I am interested in the fields of Software Development, Machine Learning and Natural Language Processing, leveraging the techniques and tools in these fields to build scalable applications. I am actively looking for full-time opportunities in these fields.
Completed Projects
Research Publication
Honors and Certifications
Experience
Data Integration Intern
June 2022 - September 2022
TigerGraph
- Developed and prototyped a Python API and JDBC driver to provide relational views for graph databases.
- Explored the use of Apache Hop as a data integration tool for building data pipelines.
- Researched, learned and documented the use of new technologies to build data connectors for graph databases.
Machine Learning Engineer
July 2020 - September 2020
Pikkal and Co - B2B Podcast Agency
- Engineered an autonomous topic generator for tagging audio podcasts in Python, delivered through API endpoints.
- Setup Postgres and Flask microservices architecture with an HTML and Javascript frontend to visualize analytics.
- Generated listener engagement charts for three different podcasts to identify strengths and weaknesses.
Backend Developer and Project Mentor
October 2017 - May 2020
DJ Unicode
- Worked on creating a book exchange platform to facilitate the easy transfer of books among students in the department View Project
- Mentored different teams of sophomores and juniors to help build projects to help students as well as faculty
Junior Data Scientist
November 2020 - April 2021
Pivony
- Setup ETL processes in Python to fetch review data from Twitter, Google Play Store and Instagram.
- Optimized and fine-tuned topic modeling algorithms LDA and GSDMM which provided KPIs to clients.
- Created 30 APIs and eliminated redundancies in code to improve performance of the software pipeline.
- Redesigned the Linux development environment in AWS, migrating from Docker to using EC2 instances.
- Automated data storage and writes to DynamoDB instead of Postgres, scaling capacity and reducing cost to company by 10%.
Machine Learning Intern
June 2018 - August 2018
Gajshield Infotech Pvt Ltd
- Prototyped an anomalous login detection system to flag suspicious system login attempts based on past user activity
Education
Masters of Science, Computer Science
2021 - 2023 (expected)
University of California, San Diego
GPA: 4.0/4.0
Relevant Courses: Design and Analysis of Algorithms, Probabilistic Reasoning, Recommender Systems, Graduate Network Systems
Teaching Assistant: Statistical NLP (SP22), Recommender Systems and Web Mining (FA22), Data Science in Practice (WI23)
Bachelor of Engineering, Computer Engineering
2016 - 2020
Dwarkadas J Sanghvi College of Engineering, Mumbai
GPA: 9.6/10.0
Relevant Courses: Algorithms, Software Engineering, Operating Systems, Computer Networks, Machine Learning
Selected as one of the top 91 Engineers in India by Economic Times (View)
Publications
Application of Deep Learning in Facial Recognition
A study and summarization of the different techniques tested and experimented so far for Facial Recognition using Deep Learning techniques.
Publication
Document Summarization using Graph Based Methodology
Abstract: The paper works towards constructing a short summary of documents with the help of natural language processing techniques. The authors goal is to identify the important aspects of a large piece of textual information, extract it and present it in a concise manner such that it conveys the information in a more efficiently and precisely. The proposed approach will generate a simple summarization of one or more documents which will help the readers to understand what the documents offer to them and identify their context without reading through them entirely. The existing methods for this work focus on different aspects of the text involved but the efficiency of these methods largely varies. The proposed methodology makes use of a combination of multiple aspects of text instead of a single aspect in order to improve the efficiency of summarization systems. The authors present a qualitative and quantitative analysis of their system as compared to the existing base-lines and demonstrate our system for a relevant application like news snippet generation.
Project
Publication
Paper
Stock Price Prediction using Grammatical Evolution
Abstract: Grammatical evolution is an evolutionary method that is used for the automated generation of programs. Over the years, different studies have proven the relevance and efficiency of this method in a wide array of fields. This method can substitute various other machine learning algorithms and older architectures to provide good efficiency and performance for optimization of algorithms. The paper aims to apply GE to predict the price of various stock market indices. An open source implementation PonyGE2 that was developed by the Natural Computing and Applications group at UCD has been employed in this paper. With the help of an objective function and a search space defined by the grammar, the evolutionary computation of the optimum solution is achieved. The effect of tweaking the grammar rules to provide different production options helped visualize the difference in the fitness of the functions generated and the consequential effect on the output produced.
Publication
Paper
Contact
ajeswani@ucsd.edu
+1 (650) 505-6339