I am a Data Science graduate student and aspiring AI Engineer, actively seeking internships and full-time roles in data science, AI engineering, and software engineering to build systems that turn messy, real-world data into decisions people can trust.
M.S. Computational Data Science @ UC Riverside — expected Dec 2026
Currently Data Analytics Intern @ Marketeq Digital (remote, through Dec 2026)
Selected for the LLNL Data Science Challenge — built an agentic pipeline for CT-based defect detection in 3D-printed lattice structures
Interested in Agentic AI, Multi-Agent LLM Pipelines, Full-Stack Data Engineering, and ML Research
Two-week intensive program for UC Riverside and UC Merced students
Designed a self-correcting registration step for aligning CT scans of 3D-printed lattice structures, addressing misalignment errors that were undermining reliable defect detection in the raw scan data.
Built an independent validation layer to cross-check the pipeline's own defect flags before reporting, reducing the risk of false positives reaching the research team.
Developed an interactive dashboard (FastAPI, Streamlit, Napari, WebSockets) so reviewers could visualize scan results and defect flags directly, instead of digging through raw output.
Presented the full pipeline as a scientific poster and 7-minute oral talk to the LLNL cohort, communicating the defect-detection approach and validation results to a technical audience.
Consolidated fragmented campaign data from Google Ads, Meta, and HubSpot into unified dashboards, replacing manual cross-platform checks and giving the marketing team a faster, single view of performance to guide targeting decisions.
Built B2B data enrichment pipelines to automatically populate missing firmographic and contact attributes, improving lead completeness and reducing manual research for the sales team.
Developed a predictive lead scoring model to rank prospects by conversion likelihood, enabling sales representatives to focus outreach on the most promising opportunities.
Applied NLP and vector embeddings to classify lead intent and surface semantically similar customer records, streamlining lead qualification and reducing manual review.
Research Assistant, Part-Time (Autoformalization / LLM Reasoning)
Riverside, CA
Analyzed the ATF (Autoformalizer with Tool Feedback) paper on Lean 4 autoformalization, identifying that its LLM-judge consistency check has only ~60% recall and degrades sharply across revision attempts (69.5% to 8.8% success by attempt 8).
Proposed a follow-up direction to build a realistic consistency benchmark from real formalizer failures using the released Numina-ATF dataset or an open formalizer like Kimina-Autoformalizer-7B.
Designed a localized, structured judge-feedback approach — flagging the specific problematic token/quantifier instead of binary pass/fail — to test whether it improves revision success on later attempts.
Scoped the proposal to student-project scale (inference-only judge evaluation, or small LoRA fine-tuning on 7-8B models).
LLM-as-JudgeAutoformalizationLean 4LoRAResearch
Nuvama Wealth and Investment LimitedJan 2025 – May 2025
Data Science Intern
Mumbai, India
Performed 40+ SQL-based negative data checks across UCC creation, registration codes, and EQ/OTC sync, driving data accuracy to ~99%+ across client buckets.
Built Python Selenium scripts to automate portfolio syncing, replacing manual PAN-by-PAN entry into the MARS platform.
Reviewed 40+ high-net-worth client portfolios in Excel to flag tail-end holdings under 1% of portfolio value, giving RMs a clear view of allocation gaps.
Analyzed synced portfolio data using Pandas, NumPy, and SciPy to assess risk and diversification, feeding into rebalancing strategies aligned to each client's risk profile.
SQLPythonSeleniumPandasNumPySciPyTableauExcel
Education
University of California, RiversideExpected Dec 2026
M.S. Computational Data Science
Coursework: Artificial Intelligence, Data Mining Techniques, Database Management Systems, Colloquium in Computer Science, DS Ethics, Foundations of Machine Learning, Data Analytics and Exploration, Natural Language Processing, Introduction to Deep Learning, Advanced Machine Learning.
Agentic pipelines, fraud models, full-stack apps, and data infrastructure.
Four focus areas: agentic AI, data engineering, applied ML, and full-stack. Click any project for the full write-up.
Hover to preview · click any node for the full write-up
Contact
I'm actively looking for full-time and internship roles in software engineering, AI engineering, and data science. Reach out — I usually reply within a day or two.
