shel. — resume

Experience

Boehringer Ingelheim

Sep 2025 — Present

Principal Applied Scientist

Own technical scoping and solution architecture for applied AI consulting engagements within the Global Animal Health division, defining conceptual frameworks and initial implementation approaches in Python, PyTorch, and TensorFlow.
Build foundational model prototypes and early-stage implementations, partnering with junior data scientists and software engineers for full development, refinement, and deployment.
Designed and implemented an LLM-based misinformation detection pipeline using Databricks and Snowflake to automate social media content triage and reduce manual review — enhancing early identification of misinformation around high-visibility products.
Delivered a production-ready proof-of-concept in 60 days, demonstrating automated content classification, clustering, and trend detection to support faster, more proactive corporate communications workflows.
Built a financial potential modeling tool to estimate purchasing capacity across major customer portfolios, enabling commercial teams to refine targeting and prioritize high-value outreach.
Lead stakeholder engagement across commercial, corporate affairs, and analytics teams, translating ambiguous business needs into scoped, actionable modeling objectives.
Navigate regulated data-access processes in a pharmaceutical environment, coordinating permissions and ensuring modeling activities comply with internal governance and regulatory standards.
Contribute to project planning, requirement definition, documentation, and cross-team communication alongside technical execution.
Mentor junior data scientists on modeling techniques, code structure, troubleshooting, and best practices for scalable, maintainable AI development.
Support ongoing improvements to analytics and AI workflows, contributing to reproducible code practices, stronger documentation, and responsible adoption of emerging generative AI techniques.

Independent Consultant

Sep 2021 — Sep 2025

Applied Scientist — Computer Vision & Health Sensing

Independent research and consulting for consumer health-sensing clients — initially alongside a full-time role at Syngenta, then full-time from 2022 — owning the full scope of ML research, measurement methodology, and analytical infrastructure for each engagement.

Led original research into computational quantification of physiological skin properties, developing novel metrics and deep learning systems to measure attributes — including radiance and tone — previously existing only as qualitative descriptors.
Developed composite scoring systems quantifying skin radiance from multi-modal physiological features, validated against expert grading.
Built a large-scale perceptual skin-tone classification system for a client, trained on a private database of 2M+ images using Vision Transformers.
Developed a CNN-based anomaly detection system for large-scale clinical imaging databases, deployed via AWS SageMaker with automated dataset quality evaluation and a client-facing SDK.
Designed and executed A/B testing and causal inference studies, owning protocol design and analysis to isolate causal effects of product features on user behavior.
Worked with real-time capacitive sensor data capturing physiological skin states, building longitudinal profiling systems modeling individual baselines over time.
Conducted customer segmentation and clustering using K-means, PCA, and UMAP to drive strategic product decisions.
Performed statistical analysis and causal inference using SQL and Python (Pandas, scikit-learn, NumPy, SciPy).
Utilized AWS Athena for large-scale dataset retrieval and processing; MongoDB/geoJSON for geolocation-based behavior analysis.
Developed interactive dashboards (Plotly Dash, Tableau) to visualize trends for stakeholder decision-making.
Defined and operationalized core product health metrics (engagement, retention, conversion).
Communicated technical findings to clients and non-technical stakeholders, translating business needs into development priorities.

Syngenta

Sep 2021 — Sep 2022

Data Scientist

Built predictive models on haplotype biological datasets to identify traits and optimize product performance based on environmental and genetic factors, combining biological domain reasoning with ML modeling end-to-end.
Designed and deployed scalable bioinformatics workflows for trait-based prediction and optimization, integrating outputs into relational databases.
Collaborated cross-functionally across the data science lifecycle — data wrangling, exploratory analysis, hypothesis testing, prototyping, validation, and deployment.
Launched a decision-making analytics platform as Product Owner on a small blended team, owning the product vision and backlog using Agile and Scrum methodologies.
Communicated complex analytical work to technical and non-technical stakeholders throughout project execution.

NC Research Campus

Sep 2020 — Sep 2021

Postdoctoral Researcher

Designed and implemented analytical pipelines for large-scale next-generation sequencing (NGS) datasets in Python across UNIX and cloud/HPC environments, establishing the computational infrastructure underlying all lab research.
Developed pipelines for transposable element polymorphism (TEP) detection and structural variant association analysis, applying GWAS-adjacent methodologies to link insertion presence/absence variations to phenotypic outcomes.
Extended NGS pipeline work into protein sequence analysis and peptide/protein structure annotation, building optimization workflows in cloud/HPC environments.
Owned computational research projects end-to-end, collaborating with graduate and undergraduate students on analytical methods and pipeline implementation.
Maintained rigorous code documentation and data stewardship practices, ensuring reproducibility and accessibility of datasets across ongoing research.

UNC Charlotte

Jan 2017 — Sep 2020

Graduate Researcher & Teaching Assistant

Conducted dissertation research on large-scale genomic annotation of the A. sativa genome, designing and deploying real-time analytical pipelines using Illumina and PacBio sequencing data in Python and Linux environments.
Extracted and integrated data from large bioinformatics databases (NCBI, GenBank), developing optimized workflows for large-scale genomic data processing.
Provided computational support for experimental workflows, collaborating with scientists to optimize data models and ensure efficient data capture.
Served as teaching assistant and independent course instructor at different points, advising students on bioinformatics tools, data analysis methodologies, and industry-relevant workflows.

Selected Projects

iridis.

Perceptual Color Analysis · 2024–

Open-data pipeline extracting robust CIE Lab/LCh color features from ~17.8K dermatology images, clustering with MiniBatchKMeans and CIEDE2000 perceptual merging. Discovered clusters are ~2.5x more predictable than clinical Fitzpatrick labels from identical features.

Read the write-up →

veridian.

Research Cognition Engine · 2025–

Corpus-grounded research cognition engine: retrieves PubMed literature, clusters and semantically summarizes it, grounds a user's own reasoning against the retrieved corpus, and builds a typed knowledge graph of clusters and entities.

Read the write-up →

topos.

Stability-First Discovery Framework · 2025–

Stability-certified protocol for deciding when latent structure in high-dimensional biological data is real: perturbation stability, matched-model comparison, and explicit go/kill criteria before expensive escalation. Applied across three organisms.

Read the write-up →

recolo.

Agent Memory Architecture · 2026–

Bio-inspired memory framework for LLM agents: episodic and semantic memory stores, tunable decay, salience-based retrieval, and a scheduled consolidation loop modeled on hippocampal replay. Extends prior semantic-clustering work as a cross-session context layer.

RepBox.

BMC Bioinformatics · Published 2023 · doi:10.1186/s12859-023-05419-5

Bioinformatics pipeline for identification and classification of novel repetitive genomic elements. Demonstrated 7% growth in detected repetitive elements and increased diversity of identified types across the A. sativa genome.

Peer-reviewed · BMC Bioinformatics 2023