📉
Price Drop Prediction System
Predicts product price drops across 1,200 products using 8 years of monthly pricing data. Improved alert precision by 3× and increased AUC from 0.43 to 0.70 over fixed threshold baselines.
LightGBMFastAPIPython
Bhoomika Gupta
_
About
Hi, I'm Bhoomika, a Data Science undergraduate at San José State University with a strong foundation in machine learning, large language models, and applied AI. Some of my recent work focuses on LLM safety evaluation, reinforcement learning-based emotion prediction, and machine learning for health and biological data. I'm especially excited about making AI systems safer, more interpretable, and useful in practice. I enjoy turning complex research ideas into systems that create real-world impact.
Research Interests
LLM Safety & EvaluationAI AgentsML for Health & Biological SystemsInterpretability
Research
Nanyang Technological UniversityActive
RL Framework for Emotion Prediction
Advisor: Prof. Erik Cambria
Developing an agentic reinforcement learning framework to anticipate speaker emotions before the next utterance using prior conversation history. Focusing on speaker-state memory, cross-speaker affect signals, and reward design for more context-aware dialogue systems.
Reinforcement LearningEmotion PredictionDialogue Systems
University of Southern CaliforniaActive
FastMap-Based Edge-Orientation Optimization
Advisor: Prof. Satish Kumar Thittamaranahalli
Building a FastMap-based framework for edge-orientation optimization under asymmetric routing demands. Modeling street and warehouse routing as a combinatorial optimization problem with directed arcs and capacity constraints.
OptimizationGraph AlgorithmsRouting
University of California, Santa Cruz · AIEA LabActive
Neuro-Symbolic AI
Advisor: Prof. Leilani Gilpin
Developing systems that integrate large language models with symbolic structures to enable structured, verifiable reasoning. Focusing on architectures that combine neural language understanding with formal logic for robust inference.
LLMsSymbolic SystemsReasoning
San José State UniversityActive
OpenGuard: Policy-Driven LLM Evaluation
Advisor: Pranoy Kovuri (AI Engineer at Apple)
Building a policy-driven framework to evaluate LLM guardrails against runtime-defined moderation rules. Designing structured test suites across attack types and analyzing failure modes such as false positives, partial leaks, policy conflicts, and degradation under constraint.
AI SafetyLLM EvaluationGuardrails
San José State University
Active2 Posters
RNA Biosensor Classification
Advisor: Dr. William Andreopoulos
Building ML pipelines for low-cost biosensors to classify RNA sequences for virus detection in water samples. Achieving 95% accuracy using a Mixture-of-Experts architecture with reproducible feature extraction.
PyTorchMoEBiosensors
San José State University · SKILLabActive
HTR-Former: Hypertension Therapy Response Prediction
Advisor: Prof. Jelena Gligorijevic
Building a time-aware transformer model to predict post-treatment blood pressure control from longitudinal EHR data. The project focuses on modeling patient history across diagnoses, medications, labs, and vitals to support therapy response prediction.
EHRTransformersClinical ML
University of California, San DiegoActive
Computational Analysis of Online Communities
Advisor: Nazanin Sabri (PhD at UCSD)
Analyzing large-scale social media data using topic modeling and sentiment analysis to map community structures and value systems across online subcultures.
NLPTopic ModelingSocial Computing
San José State University · Bachelor ThesisPlanned
Constraint Vulnerability in Chain-of-Thought Reasoning
Advisor: Dr. Teng Moh
Investigating how chain-of-thought reasoning affects instruction-following accuracy in LLMs, with a focus on identifying which constraint types are most vulnerable.
LLMsReasoningEvaluation
Experience
AI/ML Fellow
Apr 2026 – PresentBreak Through Tech
Selected for a competitive, industry-aligned AI/ML fellowship focused on applied machine learning, technical mentorship, and project-based collaboration.
MLFellowshipApplied AI
Software Engineering Intern
2026 – PresentSoftware and Computer Engineering Society · SJSU
Building self-update and restart features for SCE's CI/CD system to improve deployment reliability across internal tools.
FastAPICI/CDPythonDockerDevOps
Edge AI Intern
Apr 2026 – PresentSiMa.ai × SJSU
Developing YOLO-based edge vision pipeline for real-time occupancy estimation across video streams.
YOLOEdge AIComputer Vision
AI/ML Intern
Nov 2025 – Feb 2026EPA ESA × SJSU
Built an AI agent to validate and normalize 200+ county-level soil datasets. Automated compliance checks using rule-based and LLM-assisted evaluation, reducing manual effort by 70%.
PythonLLMsMongoDB
Research Intern
Aug 2025 – Dec 2025San José State University
Developed ML pipeline for low-cost biosensors, classifying 500+ RNA sequences with 95% accuracy using Mixture-of-Experts in PyTorch.
PyTorchML PipelineBiosensors
Fellow
Jan 2025 – Sep 2025Paragon Policy Fellowship
Analyzed large public datasets using SQL to identify trends in AI governance and healthcare. Translated quantitative findings into actionable policy briefs.
PolicySQLAI Governance
Fellow
2024AI4ALL
Built an accessibility tool simplifying complex text using FLAN-T5, reducing text length by 60–80% while preserving meaning.
NLPFLAN-T5Accessibility
Projects
🔍
Real-Time LLM Monitoring System
Streaming pipeline analyzing prompt–response events via Kafka. Handles 1K+ events/min with <200ms latency, flagging 15–20% of outputs for intervention.
KafkaStreamlitDocker
♿
AI Accessibility Assistant
Text simplification tool using FLAN-T5 for readers at varying levels. Reduced text length by 60–80% while preserving meaning across 100+ passages.
FLAN-T5NLPTTS
🧠
MyMindHealth
Mood tracking app using NLP and sentiment analysis to detect emotional patterns and surface data-driven insights.
NLPscikit-learnSQL
Leadership
Lead Ambassador
Responsible Computing Club · SJSU
Selected for a global initiative promoting ethical technology and responsible AI practices on campus.
Technical Workshops Lead
Applied Engineering Organization · SJSU
Organized and led technical workshops on AI, GenAI, Python, and Java for peers and club members.
AI Officer
AI/ML Club · SJSU
Managing and collaborating on AI projects across student teams at SJSU.
Honors & Awards
President's Scholar·San José State University·4.0 GPA
2025President's Scholar·San José State University·4.0 GPA
2024Let's Connect
Always on the lookout for opportunities to learn more and grow.