AI Systems Engineer ·
Data Science student at UC San Diego with a passion for machine learning, AI, and software development. My goal is to build meaningful solutions that leverage data to solve complex problems.
About
I am a Data Science student at UC San Diego with a passion for machine learning, AI, and software development. My goal is to create meaningful solutions that leverage data to solve complex problems.
Experience spans TensorFlow, Scikit-learn, NLP, generative AI, agentic AI, MCP-based workflows, and end-to-end full-stack delivery with React/Node. I regularly use Claude Code and Cursor for high-velocity AI engineering, and ship analytics/ML systems with cloud tooling including AWS, GCP, and Tableau-driven reporting.
Education
La Jolla, California
GPA 3.96
Expected June 2027
Honors: Provost Honors (all quarters to date)
Relevant Coursework: Data Structures & Algorithms, Statistical Methods, Linear Algebra, Calculus for Science & Engineering, Introduction to Data Science, Managing Diverse Teams, Statistical Analysis, Data Visualization, Probabilistic Modeling, Machine Learning
Experience
Most recent first.
Featured Roles
Leadership and high-impact roles highlighted first.
May 2026 - Present
UniCircle
San Diego, CA
Own all technical architecture and infrastructure for unicircle.app (React/TypeScript, Supabase, Vercel). Designed a FERPA/CCPA-compliant backend with Supabase Edge Functions and PostgreSQL. Lead engineering contributors, code review, CI/CD, and the technical roadmap with the CEO.
Nov 2025 - Present
UCSD CSE
San Diego, CA
Research with Prof. Deepak Kumar (UCSD) and Prof. Jeremy D. Foote (Purdue) on chatbot-driven interventions and Reddit user behavior. Co-authoring a paper on whether AI-mediated conversations can reduce toxic discourse at scale. Built data pipelines and evaluation workflows for generative AI and agentic AI research, including MCP-based tooling and model analysis.
Oct 2025 - Apr 2026
Engineers for Exploration (E4E)
San Diego, CA
Lead an interdisciplinary team building an AI system that predicts mangrove presence by fusing drone imagery and Sentinel-2 satellite data for Scripps Institution of Oceanography. Oversee data pipelines, feature fusion, and model integration with Python, TensorFlow, NumPy, and Google Earth Engine.
Jun 2025 - Sep 2025
PwC India
Gurugram, India
Built an agentic AI chatbot for natural language AWS/Azure/GCP onboarding across 15+ cloud services. Designed multi-agent orchestration with LangGraph, Gemini 2.5 Pro, and LangChain, cutting manual configuration time by 60%. Implemented secure credential management, automated baseline security with Azure Key Vault, and Tableau-ready reporting outputs for stakeholder visibility.
Jul 2025 - Present
DS3, Data Science Student Society
San Diego, CA
Lead ideation and selection of 10-12+ quarterly data science projects. Mentor teams through full data pipelines, including preprocessing, feature engineering, model selection, and deployment. Coordinate timelines and evaluation metrics with senior leadership for end-of-quarter showcases.
Mar 2025 - Jun 2025
Halicioglu Data Science Institute, UCSD
San Diego, CA
Conducted weekly discussion sections and office hours for DSC 40A covering empirical risk minimization, optimization, regression, classification, and discrete probability. Evaluated assignments and exams with structured feedback.
Jun 2026 - Present
CSES Open Source, TritonSpend
San Diego, CA
Lead CSES Open Source strategy, org-wide engineering initiatives, and cross-team execution while continuing to drive TritonSpend product direction and delivery.
Career Progression
Software Developer
Nov 2024 - Jul 2025
Engineering Manager
Jul 2025 - Jun 2026
President
Jun 2026 - Present
Oct 2024 - Jan 2025
Triton Quantitative Trading (TQT)
San Diego, CA
Built a hybrid LSTM/GRU forecasting model with Monte Carlo simulations (Geometric Brownian Motion). Integrated technical indicators, VADER sentiment analysis, and Sharpe Ratio risk metrics. Deployed as a modular Streamlit app with yFinance and REST API support.
Jun 2024 - Sep 2024
IIT Guwahati
Assam, India
Developed a ResNet50 CNN for SAR image classification achieving 80% accuracy across land cover types. Applied Lee and Gamma MAP speckle filters. Evaluated with precision, recall, F1-scores, and confusion matrices.
Projects
Production systems, research, and competition wins.
Spring 2026
Built an end-to-end breast tumor classification system combining WDBC SVMs and IDC image models (CNN/ViT) with FastAPI inference, explainability modules, and production deployment. Achieved 98% malignant recall, 98% precision, and 0.995 ROC-AUC.
May 2026
Reproduced the official CSE 151B Math Reasoning judger and improved Qwen3-4B-Thinking-2507 with inference-time methods (judger-aligned prompt builder, two-phase MC strategy, and no-box retry), finishing 23rd public (0.720) and 19th private (0.634) without changing base weights.
Jan 2026 - Mar 2026
Mentored full-stack environmental intelligence platform predicting invasive plant species risk. Combined map-based scanning with a hybrid ML risk engine (climate similarity, aggression traits, invasion history). Won DS3 showcase recognition.
Feb 2026
Fine-tuned ImageNet-pretrained ResNet18 with a 6-class head to predict Southern California cities from street photos. End-to-end training workflow with checkpointing and deployable inference.
Nov 2025 - Dec 2025
Implemented MCCFR (95.3% win rate), DQN, and NFSP v2 for Texas Hold'em. Built evaluation framework with confidence intervals and BB/100 metrics. GPU-accelerated training with TensorFlow/PyTorch and RLCard.
Nov 2025 - Dec 2025
92.79% AUC-ROC and 95.17% precision combining statistical features, user encoding, and SBERT embeddings, a 20.9% improvement over baseline. 85.32% AUC-ROC for cold-start users.
Skills
Contact
Open to AI engineering roles, research collaboration, and ambitious projects.