Mohammad – Python, Typescript, Golang

Interface automation platform designed to enable IT teams to automate and execute workflows through user interface interactions. The platform supports UI-driven workflow execution, allowing teams to automate repetitive operational processes, integrate applications without direct API access, and manage complex workflows across different systems. It focuses on reliable execution, workflow orchestration, and reducing manual effort in IT operations.

• Architected and built the job scheduling system in Go on top of Conductor/Orkes, including the Orkes API integration, internal job-state bookkeeping, and worker registration and dispatch;
• Designed and built the browser controller that recorded user workflows and replayed them on user-defined schedules, driving Chromium through Playwright bindings and direct CDP calls;
• Migrated the containerized agent from Docker Desktop to Podman, smoothing the user onboarding flow and improving resource utilization on customer machines.

Software platform for operating and controlling a next-generation sequencing instrument used for high-throughput genomic analysis. The project focused on developing the software components responsible for instrument control, workflow execution, data processing, and integration with the sequencing hardware, enabling reliable and automated operation of complex sequencing workflows.

• Shipped features and reliability improvements to instrument run configuration and monitoring software in Go and TypeScript, deployed across 100+ next-generation sequencing instruments in the field;
• Extended the bases2fastq and cells2stats C++ analysis tools to surface new sequencing-output metrics in their existing Parquet outputs;
• Improved instrument run workflows and monitoring capabilities to support reliable sequencing operations and faster identification of run-time issues;
• Worked across software components, interacting with sequencing hardware and downstream data-processing pipelines, ensuring new functionality was integrated consistently across the platform.

Laboratory Information Management System (LIMS) platform supporting genomic data workflows, laboratory operations, and analysis processes. The project involved building and maintaining software for managing laboratory data and integrating genomic analysis workflows. The platform also supported collaboration on variant-calling and genomic analysis capabilities, including work related to DeepVariant and large-scale sequencing data processing.

• Co-designed and ran an ablation study on DeepVariant, Google's deep-learning variant caller, investigating how restricting input channels affected calling accuracy and showing that channels with seemingly minimal information could still produce strong calling results;
• Contributed GCP Batch support to the Redun workflow engine, enabling Verily's internal genomics pipelines to run on GCP;
• Served as primary owner of sample ingestion and reporting pipelines on Airflow, processing thousands of samples per day across tens of assay types under a Tier 1 SLA;
• Built and maintained the in-house LIMS backend in Python and Go and frontend in React and Angular, serving several dozen lab scientists.

Education technology platform designed for K–12 schools to support special education, differentiated instruction, and personalized learning. The platform provides tools for educators to adapt learning materials and instructional activities to different student needs, helping schools manage individualized learning workflows and deliver more accessible, personalized educational experiences.

• Designed an analytics pipeline using AWS API Gateway, Lambda, and Redshift that ingested 10,000+ events per day and fed QuickSight dashboards and automated Excel reporting used by product and sales teams;
• Led containerization of the company's microservice stack with Docker and ECS, enabling flexible scale-up and cutting infrastructure costs by 25%;
• Shipped core features across the Toolkit and Pathways using a Vue.js frontend and Go API, used by approximately 60,000 educators.