Suryam's Portfolio

Hello there!
I’m Suryam Gupta
Data Scientist Aspirant

I am currently a Master's student in Data Analytics and Computational Social Science (DACSS) at the University of Massachusetts Amherst. I am intrigued by Data Science and the concept of finding meaningful insights and underlying patterns from data and building predictive models. I am also into Competitive Programming as I love coding, developing logic to solve problems, and using Data Structures and Algorithms to make time and space-efficient programs. I have previously worked with various Machine Learning and Natural Language Processing (NLP) projects, and I continue to explore and learn various tools and skills.

I am currently integrating LangChain and LangGraph to develop LLM-based applications, including conversational agents, retrieval-augmented generation (RAG) pipelines, and automated reasoning systems.

Other things I enjoy: Watching movies and TV shows, traveling, playing chess, reading books, and playing guitar

Skills and Tools

Programming Languages

Python

SQL

C / C++

Java

MATLAB

Software & Tools

Jupyter

Google Colab

RStudio

GitHub

Git

VS Code

LaTeX

Excel

Power BI

Quarto

AMPL

LIWC

ML & Data Science

scikit-learn

NumPy

pandas

Matplotlib

Seaborn

pyLDAvis

NLTK

spaCy

Deep Learning & NLP

TensorFlow

PyTorch

Keras

Hugging Face

BERT

GPT

LSTM

Transformers

GenAI & LLMs

LangChain LangGraph RAG Prompt Engineering CrewAI Agentic AI RLHF Multimodal AI Pydantic v2 Vector Databases

Research & Statistical Methods

Causal Inference Regression Analysis Hypothesis Testing Experimental Design Statistical Modeling Literature Review Academic Writing Inferential Statistics

Projects

Goodreads Book Analyzer: AI-Powered Reading Personality App

A live Streamlit web app that analyzes any public Goodreads profile and generates AI-powered insights: reading summaries in 5 tonal styles, genre breakdown, personality card, book recommendations, and review pattern analysis. Built with LangChain, Groq LLMs, and Pydantic v2.

A Causal Inference Analysis of Parental Education and Study Time on Academic Success

An inferential analysis on how a student's academic (study-time) and demographic (father's and mother's education level) factors (individually) affect their school performance.

A Study on Sentiment, Length, and Readability in Movie Reviews

This study analyzes how textual metrics such as Sentiment Polarity, Review Length, and Readability Score affect user engagement in Amazon Movie Reviews.

Master's Thesis - Evaluating Psychometric Features and Contextual Embeddings Potential for Mental Disorder Classification in Short Biomedical Texts

The study involves extracting psychological and psychometric features such as emotions, intensity, LIWC metrics, and contextual embeddings from short biomedical texts, and analyzing their potential to detect an early presence of either anxiety, panic, and depression in them.

Analyzing Amazon Reviews, Textual Metrics, and User’s Perceived Helpfulness

This study leverages Natural Language Processing (NLP) techniques to analyze 1M+ Amazon reviews for review length, breadth, readability, writing style, sentiment balance, and other engagement metrics to comprehend their effects on user's engagemnt and perceived helpfulness.

Linear & Integer Optimization using Python PuLP and AMPL

Worked with 16 real-life optimization problems; constructed their mathematical models, and found their optimal solution using CBC and CPLEX solver by employing AMPL software and the PuLP library of Python.

Published Works

Rajgarhia, H., Gupta, S., Mukherji, A., et al. (2025). An Evaluation Study of Hybrid Methods for Multilingual PII Detection. Poster presented at the NeurIPS 2025 Workshop on Evaluating Evolving LLM Lifecycle, San Diego, USA.
Link to Publication
Mitrović, S., Frisone, F., Gupta, S., Lucifora, C., Čarapić, D., Schillaci, C., ... & Singh, A. (2023, December). Annotating Panic in Social Media using Active Learning, Transformers and Domain Knowledge. In 2023 IEEE International Conference on Data Mining Workshops (ICDMW) (pp. 1269-1278). IEEE.
Link to Publication