This is my profileContact via CazVid
Aarnavi
Machine Learning Sciences Intern
Contact
Competencies
Demonstrated experience with Python.
Demonstrated experience with SQL.
Demonstrated experience with R.
Demonstrated experience with Java.
Demonstrated experience with Apache Spark.
Education
Summary
Machine Learning Sciences Intern open to new opportunities
Experience
Machine Learning Sciences Intern
Prototype an observational causal-measurement pipeline estimating hotel price elasticity from 68M+ bookings across 3 years of global data, targeting the randomized-holdout uplift that reserves ~11% of traffic held as zero-discount control.. Engineered a reusable PySpark + LGBM DML package – cross-fitted, Neyman-orthogonal nuisances (R² 0.85 / 0.95), a Robinson partial-linear estimator – validated via a Cinelli-Hazlett sensitivity analysis; shipped as a documented package.
Real-World Data Scientist – Contractor
Implemented a Bayesian probability model and anomaly-detection screen on 100K+ patients, analyzing 500+ disease trajectories, identifying high-risk (relative risk >1) and significant pairs (chi-square, ≥10 occurrences).. Deployed scalable ETL pipelines over 50M records of multilingual, structured and unstructured, relational healthcare data.. Architected a regex/dictionary/LLM normalization engine that cut comorbidity names ~80%, enabling NLP analytics.
LLM/RAG Intern
Optimized eco-friendly ingredients use in chemical manufacturing, improving efficiency by 10% across 5+ industries.. Developed a RAG architecture leveraging FAISS and ChemBERT embeddings, cutting query response times by 40%.
Research Analyst
Structured and analyzed ~400M employment profiles with Prof. Anastassia Fedyk to study global employment dynamics.. Engineered GPT prompts to label 100K+ unique skills at a 95% validity rate for downstream ML applications.
Quantitative Analyst
Developed trading strategies using regression, time-series analysis, and Monte Carlo simulations for portfolio optimization.
President
Led a team of 12 board members to promote math-related events, pilot outreach to alumni and professions, and collaborate with STEM clubs on events – such as Women in Quant Finance panel with Traders @ Berkeley.