Infrastructure, MLOps, AIOps, multi-cloud, database, configuration management, systems, etc.

Devops and backend engineering on a fintech/defi startup. LEPP stack, AWS, custom defi protocol engineering, fintech apis, etc.

Infrastructure, database, devops, and data engineering consulting and advising. Clients including Chive.com, Atmosphere.tv, Gather Flora. Technologies including AWS/GCP, Puppet, Saltstack, Python, custom deployment systems, etc.

Technology and team "turn-around" focusing on systems, backend, and deployment. Consolidated and eliminated unnecessary complexity in a micro-services architecture done badly with a team of ~10. Switched to a simpler chatops deployed system with a mono-repo encompassing frontend, backend, config management, and deployment tooling.Led backend team on a business pivot while building new a new database schema and graphql backed client api. Trained and mentored the team on pragmatic and simple database and application server design while balancing complexity among all parts of the organization to build a robust and efficient system that we could effectively iterate on with a small team.

Built backend infrastructure from "Initial commit" to launch of first user-facing product. Technologies used: Nginx / OpenResty: ssl termination / load balancing Flask / gunicorn: wsgi application server Mongodb: database storage Redis: cache / task queue (Python rq) AWS EC2, VPC, Route53: Hosting / Networking / DNS Saltstack: config management and deployment Supervisord: Process management Airflow / Python: Data task management and ingestion GCP / BigQuery / GCS: data warehouse NSQ: event streaming queue Metabase: Data dashboarding and visualization Fluent: Log aggregation Squid: HTTP(s) proxy Ubuntu LTS: Production deploy target OSX: Dev deploy target Buildkite: Build / CI frontend Rollbar: Exception monitoring (Python and Javascript) Server Density: Server and service health monitoringI have fairly deep knowledge of these technologies and deployed and managed all this as a one-person infra team while also delivering on various backend product features including managing the key data schemas, data ingestion and export code, timeseries data storage (tens of billions of datapoints), and in-office IT including network and switch management.OSS contributions mainly to Python libraries we used including fixing issues in requests (urllib3 https://github.com/shazow/urllib3/pull/1174), flask-classful, and a couple other smaller libraries. Led weekly engineering focused meeting looking to improve tools and processes used within the team as well as teaching engineers more about the various software, systems, and design patterns that make for building robust engineering solutions.

Reworked infrastructure hosting and deployment using AWS VPC and Ansible for a microservices based Java backend. Automated provisioning, deployment, configuration managment, and DNS. Analyze cloud costs and ways to reduce overall spend. Rearchitected loadbalancing and failover strategy. Improve overall system security leveraging best practices in AWS VPC.Leveraged infrastructure automation to improve developer environment for working on backend services and event pipeline supporting both Linux and OSX greatly reducing time for onboarding engineers.Wrote CI build and deployment automation for mobile SDKs covering Android and iOS and improved overall robustness of build system and artifact hosting.

Principal software engineer focused on infrastructure, systems, provisioning, data engineering, continuous integration, test automation, and process. Built new backend infrastructure to host and migrate the website from Google AppEngine to Amazon EC2 using open-source solutions. The AppEngine datastore was replaced with Postgres, Redis for caching, and beanstalkd for task queues. Custom components were developed in Python to maintain backwards compatibility with AppEngine in order to facilitate a smooth transition to the new backend including tools to emulate cron and task queue workers. A software shim was used to translate the GAE APIs to work with the new backend services. Deployment and configuration management was implemented with Ansible. Created a development environment for native development under OSX and Ubuntu for the new system with automated installation and very close parity to the tooling used in the production environment. Deployed and administered Jenkins for CI including Github pull request testing plugins. Developed a new ETL and data export system to support data science/engineering with the goal of centralizing data in disparate systems into a single service for custom SQL analytics using Google BigQuery. Deployed Tableau as the frontend of choice for data visualization. Gave several talks and tutorials to the engineering team in order to get them up to speed with a myriad of topics such as Unicode, systems and software architecture, and data engineering.

Senior Engineer focused on backend services, database, deployment, analytics, networking, data center, Linux system administration, et al. Heavy use of Python and a myriad of other tech: http://vazor.com/media/tech.png Architected and built service stack supporting video fingerprinting using Python/C++. Spec'd and racked server/network hardware to host this application as well as legacy applications coming from EC2. Deployed monitoring (nagios) / metrics (graphite) and other custom backend services to support this product. Rearchitected backend video streaming database from MySQL/SQLAlchemy to MongoDB with a custom object mapper. Overall this solution is much more efficient for serving video content metadata through the frontend API service while providing a convenient programming interface. Refactored various frontend/backend services which used the old MySQL backend to use the new MongoDB backend. Built EC2 deployment tools and rearchitected deployed infrastructure to be more robust and provide better system monitoring. Deployed haproxy and migrated from apache2 to nginx. Removed several single points of failure from the system by developing scalable solutions for hosting content and caching services for the connected TVs. Designed and developed beta versions of internal metadata editor, analytics dashboard with map/reduce backend, and application exception logging and reporting. These tools provice real-time content management and analytics services to the rest of the company. Migrated development environment from subversion to git including importing all projects, branches, and source code history. Deployed a gitosis server with Trac for project/issue managment.

Technical lead on a new Facebook game (http://apps.facebook.com/topfish) driving an aggressive six-week schedule with a small focused team. Key technologies used were Python, AS3, Javascript, AMF for data transport and object reflection between the server and the client, and a custom ORM between our backend storage and backend application software layer. Wrote custom tools for data analytics and logging including funnels for A/B testing the new user and payment flows. Technical lead on http://superpokepets.com launch focusing first on building a custom forum with high game integration. Later the site was extended to bring the key game experience to the destination site while also offering some features not present on the Facebook or Myspace version of the game. Developed new features in SuperPoke! focusing on user retention, engagement, and monetization though virtual goods.

Developed large portions of a web analytics application targeted at p2p CDN customers utilizing BitTorrent's DNA (http://www.bittorrent.com/dna) infrastructure. Implemented large portions of the backend services (MySQL DB, Skytide OLAP Business Intelligence server, and Apache Tomcat Java servlet for serving AJAX JSON requests for OLAP server data), middleware (SQLAlchemy ORM on top of Python MySQLdb), webapp in Pylons (Rails inspired Python MVC framework using Mako templates). Integrated/implemented frontend javascript widgets using the excellent Ext javascript library (http://extjs.com). Also wrote Hadoop (Google GFS/BigTable inspired cluster computing software) map/reduce log parser/aggregator for ingestion of data into OLAP server. Supported and maintained an internal webapp for managing BitTorrent.com's content catalog. This is also a Pylons webapp using a Python ORM (SQLObject) on top of MySQL and talking with several internal backend services (search, trackers, etc.) for changing content deployment status. Implemented features in BitTorrent.com's store for organizing TV content into collections by series and season. This required new features in the middleware/DB and scripts for migrating existing data into new DB schema. Wrote backend scripts for content metadata translation (XML, spreadsheets, and other flat-file formats to an internal standard) and ingestion as well as running backend ingestion process for the entire BitTorrent.com content catalog.

RapidChip Slice Design Engineer 06/2004 - 03/2007 Developed and refined slice design flow and key automation components (heavy use of Python and GNU Make for this) which greatly improved overall design efficiency across the slice design organization. Extreme2 family slice engineer working on executing several of the slices in the family from specification, bonding, layout, and final verification (DRC/LVS). Extreme1 family slice engineer supporting three key slices in LSI’s portfolio targeting several instances ongoing at LSI. ASIC Customer Engineer - SVDC Bay Area Team 06/2001 - 06/2004 Developed and supported a number of special-purpose software tools across the Design Center/Physical Design Center organization addressing a number of shortcomings in the LSI FlexStream toolkit. This included a new methodology and tool for addressing routing congestion in SOC ASIC designs in Python. Executed portions of a number of live designs in the design process including presales estimates, routing/timing closure, timing driven placement/resynthesis, power drop/electromigration analysis, and cross-talk analysis. Also acted as technical lead on two large ASIC projects being the main interface between LSI and the customer’s engineering staff for multi-million dollar ASIC programs. Developed and deployed Primetime SDC constraint translation tool which translates customer SDC constraints to work with LSI backend placement/optimization tools. Implemented in C/C++ using relevant netlist/technology APIs and SDC parser API..