107 lines
7.5 KiB
ReStructuredText
107 lines
7.5 KiB
ReStructuredText
Russell Ballestrini
|
||
===================
|
||
|
||
.. class:: center
|
||
|
||
**Product & Services Engineer | 860-333-3047 | russell@unturf.com**
|
||
|
||
.. class:: center
|
||
|
||
`LinkedIn <https://www.linkedin.com/in/russellballestrini>`_ | `Self-hosted GitLab <https://git.unturf.com/explore>`_ | `Personal Website <https://russell.ballestrini.net>`_ | `Portfolio <https://www.unturf.com/software>`_
|
||
|
||
Professional Summary
|
||
--------------------
|
||
|
||
Full Stack Engineer & open-source defect researcher with 15+ years scaling production systems. Currently lead `undefect.com <https://undefect.com>`_, a systematic defect research program tracking 1,254 confirmed defects across 638 open-source vendors & 60+ ecosystems, organized into 9 Mother of All Defect (MOAD) classes with patches released to public domain. Career spans site reliability (Remind, 30M users through COVID-19 distance learning), security engineering (SOC2 & FedRAMP at Cofense, DevOps at Mandiant), machine learning infrastructure (Coursemojo), & bootstrapped SaaS (Make Post Sell, Remarkbox, uncloseai, unsandbox). Proven track record executing ideas guided by vision & mission, uniting teams through effective communication, & growing product pipelines from the ground up on AWS & bare metal.
|
||
|
||
Core Competencies
|
||
-----------------
|
||
|
||
Full Stack Software Development, Physical & Cloud Infrastructure, DevOps & CI/CD, System Arch, FOSS Development, Team Leadership, Scaling, Security, and Cost Optimizations
|
||
|
||
Professional Experience
|
||
-----------------------
|
||
|
||
**Founder & Lead Researcher**
|
||
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||
**undefect.com** - Remote | 04/2026 - Present
|
||
|
||
Established a systematic defect research program across open-source software infrastructure, launched 04/01/2026.
|
||
|
||
Launched 9 Mother of All Defect (MOAD) research programs cataloging recurring structural defect patterns across 638 vendors & 60+ ecosystems. Published 1,291 UNDF advisories & 671 intel posts documenting 1,254 confirmed defects spanning algorithmic complexity (CWE-407), shared mutable state, context leaks, plaintext credential handling (CWE-312), thundering herd (CWE-362), geometric complexity, & catastrophic regex (CWE-1333). Shipped a 297× speedup patch to Mercurial's ``hg log -G`` on 04/03/2026 via the ``mercurial-devel`` mailing list. Built scanner tooling, vendor CRM, disclosure pipeline, & a public research portal driven by machine-learning agents for scan, triage, & patch generation. Patches released to public domain under a permacomputer declaration.
|
||
|
||
Tech: Python, Pelican, Caddy, Go, SQLite, Graphviz, Stripe, Anthropic Claude, OpenAI-compatible inference gateways
|
||
|
||
|
||
**Founder & Author**
|
||
^^^^^^^^^^^^^^^^^^^^
|
||
**Make Post Sell, uncloseai.com, unsandbox.com & unfirehose.com** - Remote | 01/2019 - Present
|
||
|
||
Bootstrapped a multi-tenant e-commerce platform for digital & physical product sales.
|
||
|
||
Developed & open-sourced a complete SaaS solution with a commission-free selling model for shop owners. Hot wallet payments for Monero, Dogecoin, Litecoin, & Bitcoin crypto currencies with auto quotes based on cart totals. Authored "uncloseai. Machine Learning Reference Guide to Inference Clients" using machine learning to cover inference client implementations & streaming techniques across 42 programming languages. Built unsandbox.com, an anonymous remote code compilation & execution API supporting 42+ languages with containerized isolation, artifact generation, & async execution for humans & ML agents. Built unfirehose.com, a cross-project machine-learning-agent observability platform ingesting JSONL session logs into SQLite & exposing a Next.js dashboard, todo & dependency graph system, & JSON API spanning 22+ projects (published as ``@unturf/unfirehose`` npm packages).
|
||
|
||
Tech: GNU/Linux, Python, Pyramid, SQLAlchemy, Caddy, Docker, Gitlab Runners, TypeScript, Next.js, Tailwind, better-sqlite3, Turborepo
|
||
|
||
|
||
**Full Stack Engineer**
|
||
^^^^^^^^^^^^^^^^^^^^^^^
|
||
**Coursemojo** - Remote | 10/2022 - 10/2025
|
||
|
||
Led the development & scaling of an AI assistant teacher for public schools, focusing on infrastructure management & application scaling.
|
||
|
||
Designed and implemented CI/CD pipelines for two monolithic services using AWS CloudFormation and GitHub Actions. Introduced Datadog with Fargate sidecar agents for APM, logging, & custom metrics. Scaled AI-powered educational tools using commodity language models (openai & anthropic). Implemented student monitoring tools to enhance the teacher's birds eye view of the classroom. Websocket enhancements like class pause, fastforward, & single-student restarts.
|
||
|
||
Tech: AWS (CloudFormation, ECR, ECS, ALB), GitHub Actions, Docker, Nginx, uWSGI, Python, Datadog
|
||
|
||
|
||
**Senior Security Engineer**
|
||
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||
**Cofense** - Remote | 07/2020 - 03/2022
|
||
|
||
Spearheaded security automation and compliance initiatives for SOC2 and FedRAMP environments. Managed AWS account controls & user access management for cloud. Ensured security integrity in product delivery. Gitlab, Okta, DefectDojo, CloudMapper, ECR vulnerability scans.
|
||
|
||
|
||
**Lead Site Reliability Engineer**
|
||
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|
||
**Remind** - Remote | 06/2017 - 07/2020
|
||
|
||
Managed AWS infrastructure and led cloud cost optimization, security enhancements, and developer productivity improvements.
|
||
|
||
Ensured critical services scaled during COVID-19 lockdowns, supporting distance learning initiatives. Introduced Datadog with EC2 agents for container APM, logs, & custom metrics. Infrastructure-as-code using Python & Stacker. Contributed to SlashDeploy tool for GitHub deployments via Slack commands, pioneering deployment automation before modern CI/CD platforms. Authored technical blog posts.
|
||
|
||
Tech: AWS (ECS, EC2, SQS, RDS, DynamoDB), GNU/Linux, Python, Ruby, CloudFormation, Datadog, RabbitMQ
|
||
|
||
Earlier Career Highlights
|
||
-------------------------
|
||
|
||
- **Senior System Engineer** at Mobiquity - Remote (06/2016 - 06/2017)
|
||
- **DevOps Engineer** at Mandiant - Remote (07/2014 - 05/2016)
|
||
- **Founder** at Remarkbox - Remote | 02/2016 - Present
|
||
- **System Engineer** at Cars.com - Remote (10/2012 - 06/2014)
|
||
- **Founder** of LinkPeek - Remote (01/2012 - 01/2022)
|
||
- **Unix System Administrator** at CSC - On site across 3 data centers - DoD Contractor
|
||
|
||
Education
|
||
---------
|
||
|
||
**Associate of Science in Computer Science**
|
||
|
||
Three Rivers Community College
|
||
|
||
**Bachelor of Science in Computer Science**
|
||
|
||
Eastern Connecticut State University - 93 credits towards degree
|
||
|
||
|
||
Technical Skills
|
||
----------------
|
||
|
||
- **Operating Systems**: GNU/Linux, Ubuntu, Debian, Redhat, Alpine
|
||
- **Programming**: Python, Bash, PHP, Perl, Ruby, Go, JavaScript, Java
|
||
- **Cloud**: AWS, GCP, Digital Ocean, unsandbox
|
||
- **DevOps**: Gitlab, SaltStack, Ansible, Packer, Docker, Kubernetes, Firecracker, LXD/LXC, KVM, libvirt, gitlab-ci runners
|
||
- **Databases**: Sqlite3, MySQL, PostgreSQL, DynamoDB, MongoDB
|
||
- **Queues**: SQS, RabbitMQ, ZMQ
|
||
- **Web**: Flask, Pyramid, uWSGI, Nginx, Apache, Caddy, SSE, Websockets, OpenAPI
|
||
- **Machine Learning & AI Agents**: Built opencompletion.com, a multi-user chatroom integrating with all major LLM providers (OpenAI, Anthropic Claude, Google Gemini, xAI Grok, Groq, MistralAI, Together.ai, AWS Bedrock) & self-hosted inference servers (vLLM, llama.cpp, Ollama, Unsloth). Built uncloseai.com, a free OpenAI-compatible inference gateway with reference implementations across 42 programming languages demonstrating streaming completions, tool use, & structured outputs. Built unsandbox.com as code execution infrastructure for AI agents, enabling autonomous code generation, testing, & verification in sandboxed containers. Experience with prompt engineering & agentic workflows orchestrating multi-step reasoning & tool calling.
|