Resume

Ashley Taken

Machine Learning Engineer · Data Scientist · Applied AI

Professional Summary

Machine Learning Engineer and Data Scientist with 6+ years of experience developing, deploying, and operating production ML systems. Expertise spanning deep learning, reinforcement learning, quantitative research, and LLM-based systems, with an emphasis on rigorous experimentation and end-to-end engineering ownership.

Education

M.S. Analytics · Computational Track
Georgia Institute of Technology · 2025
B.S. Chemical & Biomolecular Engineering
University of Houston · 2017
B.S. Biology · Minor in Psychology
Texas A&M University · 2014

Publications

Automated Well Status Identification in US Offshore Operations Using Machine Learning
SPE Offshore Europe · 2025 · Co-Author

Developed XGBoost and 2D CNN models with engineered features to improve offshore well status classification accuracy.

View Publication

Experience

Occidental Petroleum (Oxy)
Data Scientist
2024 – Present
Contract · Remote
  • Deployed Dockerized 2D CNN inference services in Azure, improving real-time well-status prediction accuracy from 77% to 98%.
  • Built full-stack engineering applications using Python, Flask, Dash, SQL, Docker, Azure DevOps, and AVEVA PI.
  • Architected a 22-agent enterprise RAG ecosystem across multiple engineering domains.
  • Created and maintained 25+ production manage-by-exception workflows across Seeq, Power Automate, Azure, and AVEVA PI for offshore operations.
PythonPyTorchAzureDockerAVEVA PIRAG
Taken Group LLC
Founder / Machine Learning Engineer
2024 – Present
Independent R&D
  • Built a full-stack quantitative ML research platform using FastAPI, SvelteKit, MariaDB, Redis, Docker, and Python.
  • Developed reproducible statistical research workflows for sampling, feature engineering, leakage-aware validation, and model evaluation.
  • Built an agentic research platform with hybrid retrieval, reranking, MCP tooling, and a dedicated evaluation framework.
  • Developed transformer and contrastive-learning workflows for financial time-series representation learning.
FastAPISvelteKitPyTorchAgentsTime Series
Imubit
Implementation Engineer (Forward Deployed Engineer)
2022 – 2024
Industrial AI
  • Designed and deployed five closed-loop reinforcement-learning systems from POC to production across industrial process applications.
  • Analyzed model behavior and supported production tuning using SHAP, hyperparameter optimization, YAML deployment configuration, Grafana, and cloud infrastructure.
  • Delivered projects producing more than $2M in documented value.
Reinforcement LearningSHAPAWSGrafana
Chevron Phillips Chemical
Process / Operations / Advanced Process Control Engineer
2020 – 2022
Chemical Engineering
  • Developed and supported advanced process control and real-time optimization applications across olefins operations.
  • Automated real-time furnace monitoring with Python by generating custom status signals and operator alerts to improve response time.
  • Modernized plant data analysis using Python regression models for distillation quality control and production optimization.
  • Built real-time and forecasted emissions dashboards using Python and Power BI to help prevent EPA compliance violations across three plants.
  • Identified approximately $350K/year in nitrogen savings through condition-based analysis of a closed-loop regeneration system.
PythonAPCSeeqPower BIProcess Engineering

Technical Skills

Machine Learning
PyTorchXGBoostTransformersCNNsLSTMsTime-Series ForecastingAnomaly DetectionRepresentation Learning
Applied AI
RAGAgentic AIHybrid RetrievalBM25Cross-Encoder RerankingMCPLLM Evaluation
Engineering
PythonSQLFastAPIFlaskSvelteKitDockerAzureAWSRedisMariaDB

Education

M.S. Analytics · Computational Track
Georgia Institute of Technology · 2025
B.S. Chemical & Biomolecular Engineering
University of Houston · 2017
B.S. Biology · Minor in Psychology
Texas A&M University · 2014

Publications

Automated Well Status Identification in US Offshore Operations Using Machine Learning
SPE Offshore Europe · 2025 · Co-Author

Developed XGBoost and 2D CNN models with engineered features to improve offshore well status classification accuracy.

View Publication