Nathan Arias
Los Angeles,CA
Summary
Applied data scientist bridging AI and life sciences by building real-world data and machine learning systems and teaching scientists how to build and use them. Focused on making data pipelines, modeling workflows, and modern AI tools accessible through hands-on training and practical applications.
Overview
4
4
years of professional experience
Work History
Data Scientist
Amgen
Thousand Oaks, California
04.2024 - Current
- Designed and implemented production-grade data pipelines that aggregate experimental data across the organization, transforming raw laboratory source data into structured, analysis-ready datasets used by hundreds of users and downstream assets such as dashboards, analytics tools, and AI systems.
- Developed predictive models for clone selection in early-stage drug development, combining deep learning convolutional neural networks and classical machine learning approaches.
- Developed an LLM-based retrieval-augmented generation (RAG) system to extract structured insights from scientific PDFs, eliminating manual transcription and enabling scalable data extraction workflows
- Developed a time-series data compression algorithm achieving up to 95% reduction in data size while preserving signal integrity for large-scale analysis
- Built automated ingestion and transformation workflows (Databricks), including deduplication, schema standardization, and validation to improve data reliability and reproducibility
- Led technical training sessions for ~30 data scientists across multiple functional teams, teaching applied AI workflows and modern data platforms including Streamlit, Databricks, Langchain, and Codex for data science upskilling.
- Authored documentation and best practice guides to support onboarding and promote self-service learning
Data and Bioprocessing Scientist
Omeat
01.2023 - 02.2024
- - Led a skilled team of five scientists in scaling bioprocessing operations critical to cultivated meat production at Omeat, utilizing design of experiments and process optimization to manufacture innovative three-dimensional cellular scaffolds, successfully piloting at 1000L scale.
- Co-inventor on a patent submission for an innovative bovine serum processing technique currently being developed into a commercial product.
- Orchestrated the development of a comprehensive relational database to systematically track and analyze bovine health indicators, leveraging SQL for data storage and complex data aggregation.
- Developed quantitative laboratory applications for monitoring microcarrier cell confluency and particle size distributions, utilizing image processing algorithms
- Developed and trained deep learning models to forecast critical blood cell counts from multifaceted data inputs such as cow health indicators and plasma donation schedules, driving improvements in animal welfare and optimizing donation cycles.
- Applied bioinformatics expertise in the analysis of proteomic and metabolomic data, enhancing understandings of novel products and bioprocessing parameters through the elucidation of key protein interactions, pathways and gene ontologies.
- Engineered interactive, real-time data visualization dashboards using Streamlit and Tableau, integrating Python scripting for advanced analytics, which significantly improved operational decision-making and strategic planning based on live data feeds.
Research Associate (Chemistry)
Beyond Meat
07.2022 - 10.2022
- - Employed advanced analytical techniques, including UV-VIS spectroscopy, High-Performance Liquid Chromatography (HPLC), and Liquid Chromatography-Mass Spectrometry (LCMS), for the isolation and detailed characterization of novel color compounds.
- Engineered high-throughput screening protocols on robotic platforms for efficient colorant selection, prioritizing candidates based on their performance in comprehensive shelf life studies and matrix compatibility, ensuring durability and sensory integrity.
- Conducted methodical bench and pilot-scale experiments, refining colorant dosing strategies to maximize color stability and cost-effectiveness while adhering to regulatory standards.
- Developed advanced web tools for CIELAB color visualization, enabling detailed spatial analysis of color shifts, and directly correlating these findings with sensory analysis outcomes to refine product formulations for optimal consumer acceptance.
- Conducted plant-based colorant research, emphasizing matrix chemistry to ensure integration without affecting food product taste or texture, while conducting shelf life studies to assess long-term color stability across various matrices.
Education
Masters - Information and Data Science
University of California, Berkeley
05-2025
B.A. - Molecular and Cellular Biology
University of California, Berkeley
11-2021
Skills
- Machine learning
- Data analysis
- Python programming
- SQL databases
- Team leadership
- Effective communication
Projects
MedCoderAI — UC Berkeley Capstone, 04/24 - 09/24, Led development of state of the art AI-driven medical coding system using a mixture-of-experts longformer model trained on the MIMIC-IV dataset. Built preprocessing pipelines, embeddings for ICD-10, and real-time integration tools for medical coders.
Languages
- Spanish
- English
Languages
Spanish
Limited
Timeline
Data Scientist
Amgen
04.2024 - Current
Data and Bioprocessing Scientist
Omeat
01.2023 - 02.2024
Research Associate (Chemistry)
Beyond Meat
07.2022 - 10.2022
Masters - Information and Data Science
University of California, Berkeley
B.A. - Molecular and Cellular Biology
University of California, Berkeley