Deploying custom large protein language models to production along with MLOps and infra (K8s, Terraform, building pipelines, multi-cloud compute workloads)

Engreitz Lab

• Built pipelines to run thousands of large biological datasets through our statistical and ML models

• Aided in publication: Building an encyclopedia of enhancer-gene regulatory elements in humans (https://www.biorxiv.org/content/10.1101/2023.11.09.563812v1)

• Experimented and built ML models to predict genomic data tracks

• Revamped the codebase in terms of usability and scalability

• Analyzed genomic datasets (RNA-seq, ChIP-seq, ATAC-seq, HiC), helping diagnose predictive models

Platform infrastructure for modeling and storing Meta's network (network devices, ip addresses, etc)

• Managed majority of performance work for the core database and services

• Led reliability initiatives around critical components

• Mentored 4 interns and 2 full time engineers

• Architected and rolled out a server-side cache and a pub/sub system

• Designed temporal datastores to enable historical and future queries of the network

Fluent is a declarative programming framework for building distributed systems

Researched different consistency models and how they fit into Fluent

Implemented a 2PC coordination algorithm for Fluent’s replication and partitioning scheme