John Bates Email & Phone Number

South San Francisco, California, United States

Verily Workbench Technical Solutions Engineer | 8/2021 - Present- End-to-end cloud infrastructure solutions for academic and industry researchers- Genomics and data science implementations and scaling for research partnersSearch and Assistant Growth Factors | 6/2020 - 8/2021- Developed machine learning models and data pipelines to drive search user growth and retention.Ultra-wide Field Diabetic Retinopathy Diagnosis | 4/2017 - 6/2020- Technical lead for the effort to diagnose Diabetic Retinopathy (DR) from graded retinal images.- Developed ETL pipelines, CNN-based deep learning models, statistical analysis tools.- Rapid iteration on model architecture, hyperparameter tuning, and image grading strategies.- Extensive cross-functional collaboration with external data partners, physicians, statisticians, and infrastructure engineers."Machine Learning Ninja" Rotation | 10/2016 - 4/2017- 6 month rotation with the Google Research and Machine Intelligence group.- Developed TFX estimator for state saving recurrent neural networks.- Wrote CUDA implementations for TensorFlow operations.Cloud Genomics | 7/2013 - 10/2016- Wrote the first implementations of key components of Google's Cloud Genomics API and backend.- These services host and process genomic data for hundreds of customers

San Francisco, Ca

Developed crop yield models based on weather data, using Clojure, Cascalog, and Mahout to run at scale on Amazon's Elastic Map Reduce service.

Berkeley, Ca

Wrote pipeline for automated discovery of CRISPR elements for all prokaryotic genomes in MicrobesOnline (MOL) Wrote data aggregation and resolution scheme for MOL, integrating metabolite, reaction, and pathway data from a variety of online databases (e.g. KEGG, MetaCyc, and BiGG.) Wrote a web-based front end for this metabolic pathway data on MOL. Writing the Genome-Linked Application for Metabolic Maps (GLAMM) - a platform for visualizing functional data (e.g. mRNA expression, protein levels, metabolic flux, metabolite levels, etc.) in the context of metabolic pathways, as well as finding routes for novel synthetic pathways for biofuel production, available at http://glamm.lbl.gov.

Cupertino, Ca

Designed and wrote large sections of BezelServices - Mac OS Xs HID event handling service. Greatly expanded the infrared receiver driver, including integrating it into the AppleTV. Wrote a variety of demo applications for prototype iPad hardware and software, and a wide variety of novel human interface technologies. Managed a small software and hardware prototyping team. Extensive cross-team collaboration with software and hardware engineering, industrial design, user interface, product design, and art teams.

Implemented and tested components of the Java JSR-184 Mobile 3D graphics API on the Imageon 2300 SOC. Integrated this API into Win32 and WinCE platforms. Coordinated with customers to make systemic changes to our existing OpenGL ES driver (written in C) to optimize performance on and add features to their handheld products.

Cupertino, Ca

Developed a transaction level simulator (transactor) in Vera to verify the behavior of a PCI Express implementation. Wrote Altivec-compatible assembly code in order to test video processing cores on a custom ASIC. Wrote and debugged transactors and tests for the HyperTransport bus implementation in the Apple PowerMac G5.

Implemented fixed-function shader components (e.g. noise generators, texture filters, etc.) in a functional, bit-accurate C simulation of custom shader hardware. Wrote and assessed the impact of hand coded, vectorized linear algebra routines using the Intel SSE2 assembly instruction set.

San Jose, Ca

Developed a tool for quantifying the correlation between circuits simulated using static timing models and those simulated using SPICE models. Wrote a tool to optimize the routing and buffering of clock lines between hardware blocks on an ASIC. Developed a tool and methodology for distributed synthesis of Verilog code, reducing the time for synthesis by a factor of four.

Bs, Electrical And Computer Engineering

Education record

Coursera: - Mathematical Biostatistics Bootcamp 1 - Machine Learning Stanford Online: - Statistical Learning