I started this project while consulting with Praxis AI, offering engineering expertise and experience building applications with Generative AI. I developed the Generative Intelligence Engine, which serves as Pria’s backend Flask server. I was responsible for creating and managing Gitlab CI/CD pipelines, deploying the application to a containerized environment in AWS, and deploying the production application to a dynamically-scaled and load-balanced Elastic Container Service.While building this generative AI backend for Pria, I learned how to utilize both OpenAI’s LLMs and open-source models via HuggingFace, worked with Langchain agents and document integration chains, and created custom vector-embedding tools to expand Pria’s knowledgebase. Pria is now being successfully used by dozens of academic institutions across the country and received two recognitions in the 2023 American Business Awards: GOLD for Achievement in Online Training, and BRONZE for Use of Artificial Intelligence in EdTech.

At Flagship Labs 84, an early-stage pharmaceutical startup, I played a crucial role as the first computational biologist tasked with developing advanced computational pipelines and managing cloud infrastructure for single-cell biomedical research.My responsibilities included architecting and executing sophisticated data processing pipelines, specifically designed for high-throughput single-cell genomics. Utilizing my Python skills and the Scanpy package, combined with expertise in AWS cloud architecture, I automated and refined numerous stages of data analysis, ensuring the efficient processing and insightful interpretation of extensive genomic datasets. My expertise was instrumental in configuring and maintaining high-performance cloud-based computational resources appropriate for the computational tasks encountered in cutting-edge biomedical research.

Managed the integration of bulk RNAseq processing tools (STAR, Kallisto, DESeq2, etc) into a single pipeline which could be deployed on DNAnexus cloud resources. The aim has been to provide a simple solution for pharmaceutical companies and bioinformaticians to process large quantities of transcriptomic data, enabling new discoveries into the mechanisms of gene expression and development of novel therapeutics.

Worked to scale bioinformatics pipelines to HPC clusters and democratized Kubernetes platforms. Customized Docker containers and wrote Nextflow pipeline scripts to perform terabyte-scale computational analysis of genomic data on the Texas Advanced Computing Center and the Pacific Research Platform using Kubernetes. Helped to provide resource usage data and documentation while testing scalability onto cloud compute resources. Throughout the course of this work, I co-first authored a publication to Cell Press along with affiliate authorship contribution to an additional four publications (https://reedbender.com/research).

Processed patient-specific RNA-sequencing fastq files alongside TCGA and GTEx datasets to create comprehensive gene expression matrices for downstream analysis.Performed differential gene expression analysis, cellular state characterization through functional enrichment, and drug target identification.Customized deep learning tools, repurposing generative adversarial networks to identify differentially expressed genes in a patient's tumor relative to public TCGA data.