Based in the Bay āļø
PDF Version š§³
Iām a Product-Minded Engineer, whoās core focus is on enhancing end-user experiences and ensuring scalability. My commitment to making a positive impact in the world is at the forefront of my work, especially through my recent endeavors in developing a SaaS-based generative AI application. This pursuit has not only highlighted the immense potential of generative AI in automating tasks and fostering creative solutions but also reinforced my belief in its role as a catalyst for innovation. My journey has taken me from collaborating in large teams at industry giants like Pinterest to implementing ML algorithms on the field in the MLB to spearheading data initiatives as the key data personnel in my current role. I thrive in fast-paced iteration and seamless collaboration, always aiming to bridge the gap between technical excellence and impactful solutions.
Work History
EchoAI
Machine Learning Engineer Lead ā¢ Feb. 2022 - Present
- Developed and implemented generative algorithms via LLMs from 0ā1, ensuring autonomous, precise, streamlined insights generation through cutting edge research, RLHF, LangChain, and vector databases
- Directed technical projects autonomously, meticulously scoping requirements and sequencing work for the engineering team, and serving as the primary LLM and ML expert advisor to the CTO and team
- Contributed to the creation of an external service using AWS Lambda for large-scale machine learning models. This effort focused on offloading computational tasks to enhance web app performance, and involved the integration of pgvector as the embedding database for optimal LLM accuracy and efficiency
- Hosted and fine-tuned open source LLMs via LoRA and DSPy for cost reduction and flexibility on task routing. Utilized a dataset, both human and synthetic, that was aligned with customer intent
- Collaborated with teams across departments leveraging dbt and Hex for advanced analytics and automated the processing of customer data requests, boosting satisfaction as the sole Data personnel at the company
eHealth
Data Scientist ā¢ July 2020 - Feb. 2022
- Identified customers likely to churn through a Random Forest classifier and MLflow for model selection, which resulted in a high AUC and recall. Results were made accessible via an ETL pipeline to Snowflake
- Determined lifetime value (LTV) for customers using a Cox Proportional Hazard model. Delivered results via AWS SageMaker and Lambda to create a REST API. This work enhanced performance marketing efforts and increased understanding of revenue flow across the organization
- Created an end-to-end model, pipeline, and monitoring to predict the LTV associated with sending a mailer to an individual via Spark, XGBoost and H2O which led to over 80% LTV with a 50% reduction in cost
New York Mets
Data Science Intern ā¢ October 2019 - June 2020
- Maximized outs using statistical distributions, XGBoost, and K-means to create an infield and outfield defensive shift model in an accessible dashboard hosted on AWS
- Enhanced organizationās understanding of fieldersā ability with a novel internal metric, plus/minus, created with a Random Forest classifier, cross entropy, and data aggregation in pandas
Business Analyst ā¢ September 2018 - July 2019
- Created a Tableau dashboard to influence decision making and improve process efficiency. Decreased the duration of time required to solve outstanding service tickets by 30%
Skills & Proficiencies
Expert
Python
PyTorch
ML Ops
MLflow
LLMs
AWS Lambda
Postgres
SQL
dbt
Advanced
React
Java
Docker
NextJS
Spark
TensorFlow
Education
University of California, Berkeley, Class of 2017
B.A., Applied Mathematics
B.A., Statistics
University of San Francisco, Class of 2020
M.S., Data Science