My Portfolio

About Me

I’m a current grad student at UCSC, training both machine learning models and myself in the art of NLP. With a Computer Science degree in AI & ML, I navigate the world of software and data—turning structured code into intelligent conversation.

I'm a passionate about creating innovative solutions to complex problems. I strive to write clean, efficient, and maintainable code. I'm always eager to learn new technologies and stay up-to-date with the latest industry trends. When I'm not dabbling with code, I like exploring new places, meeting people and connecting with nature.

Technical Skills

Python Java C/C++ Machine Learning (ML) Algorithms Deep Learning (DL) Language Modeling Computer Vision Conversational Agent systems Causal Inference MySQL MongoDB AWS GCP PostgreSQL Git Docker DSA OOPS

Education

Master of Science in Natural Language Processing (NLP)

University of California, Santa Cruz

2024 - Ongoing

Relevant coursework: Coversational Agents, Machine Translation, Advanced topics in Deep Learning

CGPA: 3.8

Bachelor of Technology in Computer Science & Enginering, specializing in AI & ML

Keshav Memorial Institute of Technology, Hyderabad

2020 - 2024

Relevant coursework: Data Structures, Algorithms, Database Management, Web Development, Software Engineering

CGPA: 3.8

Higher Secondary Education

Sri Chaitanya Junior Kalasala, Hyderabad

2018-2020

Science Stream with Mathematics, Physics and Chemistry

Percentage: 92%

Professional Experience

Student Researcher

Adobe

May 2025- Present

Working on a Multilingual NL to SQL System as part of a capstone project
Designed and built a pipeline for data retrieval using SQL for multilingual input
Modifying methods to generate a synthetic dataset for model training and evaluation
Creating a framework/model for translation and retrieval of data

Researcher and member

University of California - Santa Cruz

March 2025 - October 2025

Worked in the AIEA Lab under Dr. Leilani Gilpin (ex-Sony)
Developed a custom autonomous vehicle simulation environment on top of the CARLA simulator integrated with OpenAI’s Gym interface.
Engineered vehicle movement models and optimized obstruction-avoidance accuracy to 90%.
Analyzed and compared reinforcement learning algorithms to enhance system robustness.
Presented research findings and a technical poster at the UCSC Research Symposium (May 2025) titled “Comparing Reinforcement Learning Algorithms for Autonomous Vehicles.”

View Poster

Student Researcher

University of California - Santa Cruz

October 2024 - December 2024

Built a Multilingual fact-checked claim retrieval system under the guidance of Dr. Amita Misra (Applied Science Head, Amazon) and designed and implemented a multilingual system to retrieve top-k relevant fact-checks for social media posts.
Modified and evaluated multiple models including Logistic Regression, BiLSTM with Attention, and XLM- RoBERTa to improve cross-lingual retrieval accuracy to 0.87.
Hyperparameter-tuned XLM- RoBERTa, achieving the highest standalone model accuracy of 0.78.
Collaborated with a cross-disciplinary team to refine the fact-checking pipeline and streamline data preprocessing for multilingual datasets.

View Code

Projects

LLM-as-a-Judge

Engineered an all-to-all framework where LLMs acted as both generators and evaluators.
Also curated ELI5 and WritingPrompts datasets to test factual QA and creative writing outputs. Deployed 10 LLMs to score peers on coherence, relevance, creativity, and style.
The project revealed key trends in judge model behavior and helped identify biases like verbosity preference and self-enhancement, while also comparing LLM assessments with human evaluations.

Python LLM Dataset curation

View Code

Multilingual Fact-Checked Claim Retrieval

Developed a system capable of retrieving the top k-most relevant fact-checks for social media posts in various languages.
Created and modified logistic regression, BiLSTM with attention and XLM-RoBERTa models to retrieve relevant fact-checks
Hyperparameter-tuned XLM- RoBERTa, achieving the highest standalone model accuracy of 0.78.
Collaborated with a cross-disciplinary team to refine the fact-checking pipeline and streamline data preprocessing for multilingual datasets.

Python DL algorithms Pre-trained models NLP

View Code

Slot Tagging on Movie dataset

Trained multiple models(MLP,LSTM,CNN) to tag slots in natural language utterances of users to accomplish their requests.
Experimented with different embeddings and optimizers to find the best-fit model.
Achieved an accuracy of 89% with a BiLSTM model.

DL pre-trained models Python NLP

View Code

Inventory Management System

Achieved 0.93 accuracy, improving performance by 14.5% over YOLOv6 and 13.4% over YOLOv7.
Built and augmented a 200-image dataset to enhance object detection precision.

Python Computer Vision Flask SQLite

View Code

Image Encryption using Blowfish Algorithm

A cyber-security based web application to ensure safe encryption and decryption of images.
Developed a learning prototype to enable secure transmission of encrypted images, utilizing blowfish encryption and decryption along with an interactive user interface.

Java HTML5 CSS3 Java Servlet pages

View Code

Hey, I'm Shriya Sravani Y

About Me

Technical Skills

Education

Master of Science in Natural Language Processing (NLP)

Bachelor of Technology in Computer Science & Enginering, specializing in AI & ML

Higher Secondary Education

Professional Experience

Student Researcher

Researcher and member

Student Researcher

Projects

LLM-as-a-Judge

Multilingual Fact-Checked Claim Retrieval

Slot Tagging on Movie dataset

Inventory Management System

Image Encryption using Blowfish Algorithm

Certifications

Gen AI Job Simulation

Transformer-based NLP

Strategic Thinking

Transformer and BERT model with Google Cloud

Introducing Generative AI with AWS

Foundations: Data, Data, Everywhere

Resume

Get In Touch

Let's Connect!