Rose Howell - Software Engineer

About Rose Howell

Software engineer working across the stack, remote since 2017. I'm currently at Hyperthesis building Cognetto, a platform that pulls structured data out of scientific papers using LLMs so researchers can quickly run meta-analyses. Before that, seven years at eyeo (Adblock Plus): two on the backend team, then five as a founding member of the machine learning team that launched the first ML-powered ad blocker.

A given week might include a Postgres migration, the FastAPI routes on top of it, the React components that render the result, and the GCP infrastructure the whole thing deploys onto. I stood up our production environment this year, from an empty cloud project through to the live app on its own domain, and I own the release pipeline into it.

The work I like most is the kind where correctness is checkable: scoring that tells us whether an LLM's extraction is actually right, and highlighting that points back to the exact table cell a number was pulled from. I care about code quality, and I do my best work in async, remote-first environments with a collaborative team and thoughtful code review.

I hold a Master of Science in Computer Science & Engineering from the University of Michigan, where I worked with J. Alex Halderman on security research, and a Bachelor of Science in Information Assurance from Eastern Michigan University.

Resume

Experience

5/2026 - Present	Software Engineer Hyperthesis (Cognetto) Remote Stood up the company's production environment end to end: GCP project, Cloud SQL Postgres 17, Cloud Run services, Artifact Registry, Secret Manager, and GCS storage, plus the load balancer and SSL for app.cognetto.ai, and own every release to it since Replaced branch-merge promotion with a scheduled tag-based release pipeline in GitHub Actions and Cloud Build, gated on green builds and verified by automated post-deploy health probes Ship full-stack features across a Python/FastAPI and PostgreSQL backend and a Next.js/React frontend Improved the existing PDF highlighting so it works for values pulled from tables, anchoring each one to the cell it came from when the model returns usable coordinates and falling back to a coarser match when it does not Extended the LLM evaluation harness with recall and precision scoring for correlation effects and group statistics, giving the team a measured accuracy baseline for prompt and model changes Built the repository's first CI (pytest, Jest, shell tests) and per-branch Docker stacks for isolated full-app environments
1/2025 - 5/2026	Software Engineer Dahlke Tech GmbH Remote Contributing to software projects for climate tech, property management, and other various clients, implementing features and enhancing existing codebases using Python, Django, PostgreSQL, Docker, and other technologies as needed Refactored climate tech VC investment platform's data enrichment system from hardcoded model calls to pluggable architecture, enabling flexible integration of 25+ AI models (OpenAI, Anthropic, Google) for analyzing 8,900+ startups Designed and implemented benchmarking dashboard to evaluate and compare performance metrics across AI models, enabling data-driven model selection for investment research Built automated data synchronization pipeline using Django webhooks and Airtable API to export PostgreSQL data, streamlining client reporting workflows Working autonomously in fully-remote, asynchronous environment with minimal oversight, submitting code through pull requests and participating in code reviews
5/2017 - 4/2024	eyeo GmbH (Adblock Plus) Remote (Cologne, Germany) · Seven years, two teams Software Engineer – Machine Learning Team · 9/2019 – 4/2024 Founding member of the ML Team; helped define structure and initiatives to launch the first ML-powered ad blocker Developed and maintained browser extension to scrape data from websites, convert to graph nodes and edges, and monitor model performance in real-time (JavaScript, HTML/CSS, TensorFlow.js) Maintained backend server to accept, sort, and store incoming data, then sync with GCP cloud storage (Python, Flask, Docker, Redis, GDPR compliance) Created ML pipelines to collect, convert, and extract features from datasets, balance and enhance training data, and train/deploy new models (PyTorch, Spektral, Google Cloud AI Platform) Built comprehensive Pythonic test suites for new and existing code using Pytest, tox, Poetry, achieving 85%+ code coverage and implementing GitLab CI pipelines Developed and maintained Ways of Working and Code Quality standards, automating code quality checks Led ML Community of Practice, conducted literature survey of ML in ad blocking, and shared latest research findings company-wide Wrote and reviewed code as part of a 9 person, fully-remote, cross-functional team spanning 8 time zones Software Engineer – Backend Team · 5/2017 – 9/2019 Joined as a full-time summer intern onsite in Cologne, continued part-time through the final year of my master's, then moved to full-time after graduation Developed tools and processes for generation and publication of filter lists for Adblock Plus and website maintenance Maintained Python version of Adblock Plus, keeping feature parity with JavaScript version across multiple releases Maintained content management system (CMS) using Python and Jinja2 templates Maintained build scripts, site scripts, and other backend infrastructure code
9/2016 - 4/2018	Graduate Student Instructor University of Michigan Ann Arbor, MI Taught undergraduate computer science courses, leading weekly hands-on lab sections and holding office hours to help students with concepts and projects EECS 388 (Introduction to Computer Security): Taught cryptography, web security, network security, application security, and digital forensics EECS 280 (Programming and Intro Data Structures): Taught data structures and algorithms in C++ and Python EECS 183 (Elementary Programming Concepts): Taught C++ and Python fundamentals
5/2016 - 4/2018	Graduate Student Research Assistant University of Michigan - Security and Privacy Research Group Ann Arbor, MI Computer security research with J. Alex Halderman Researched email security in foreign countries, testing which mail servers were disallowing encryption (STARTTLS) country-wide Tested security and privacy of RFID systems with various readers Helped automate security tests in medical networks by creating virtual hospitals for use as sandboxes, honeypots, and testbeds
2/2014 - 8/2015	Technical Support Representative Barracuda Networks, Inc. Ann Arbor, MI Deployed, configured, and maintained firewalls in a variety of network environments for enterprise clients Provided technical support for firewall and email spam filter products via phone, email, and IM Troubleshot complex network issues including routing, VPN, and security configurations

Education

9/2015 - 5/2018	Master of Science in Computer Science and Engineering University of Michigan Ann Arbor, MI Magna Cum Laude
1/2014 - 4/2015	Bachelor of Science in Information Assurance Eastern Michigan University Ypsilanti, MI Summa Cum Laude
9/2010 - 12/2013	Associate of Applied Science in Computer Information Systems - Cybersecurity Oakland Community College Royal Oak, MI Summa Cum Laude

Technical Skills

Primary Language	Python (8+ years professional experience)
Additional Languages	JavaScript, SQL, HTML/CSS, Shell/Bash, C/C++
Backend	FastAPI, Django, Flask, PostgreSQL, Redis
Frontend	React, Next.js
Cloud & Infrastructure	Google Cloud Platform (Cloud Run, Cloud SQL, Cloud Build, Artifact Registry, Secret Manager, Vertex AI, Document AI, load balancing, SSL), Docker, Linux
ML & AI	LLM evaluation and prompt engineering, recall/precision measurement, PyTorch, TensorFlow.js
Testing & CI/CD	pytest, Jest, Playwright, tox, GitHub Actions, GitLab CI

Certifications

Certified ScrumMaster®
Scrum Alliance® certified, March 2020

Volunteer Work and Community Engagement

Block Club President
Co-founder and President, Detroit Block Club Program, May 2023 - Present
Phi Kappa Phi Honor Society
Member, Phi Kappa Phi #154, Eastern Michigan University, 2014-2015
Information Assurance Student Association
Secretary, IASA, Eastern Michigan University, 2014-2015
Michigan Collegiate Cyber Defense Network
Competitor, Michigan CCDN, Davenport University, March 2015
Cyberspace Security for Everyone
Panelist, 35th Annual Southeast Michigan Mensa Regional Gathering (SEMMantics35), May 2014
Oakland Community College IT Club
Member, Information Technologies Club, Oakland Community College, Royal Oak, 2010-2013