2025 Brag Document
TL;DR
Another big year of change and contrast. Worked three separate jobs. Used Teams, Slack, and Discord. Spent 8 months working remotely & 4 in person. Developed with autonomous naval software. Built AI focused developer tooling. And now I’m deep in AI memory infrastructure. First official calendar year as a working woman!
Oh also, I moved to NYC.
Goals for 2025
Get a new job!
Work closer with AI
Leave defense contracting
Move to NYC
Goals for 2026
Work hard and keep shipping things I’m proud of
Pick up consulting or freelance work on the side (Make steps toward more flexibility and independence)
Maintain work boundaries
Learn Spanish
Publish more articles on blog
Projects
Leidos - Maritime (Jan)
Final month on the team before transitioning
- Planned work roadmap for the new year
- Package management preparations (rpm)
- Owned development of a maneuver in the autonomous software and the Human Machine Interface (HMI)
- Attended an NVIDIA conference on developing with their AI workflows
Leidos - Software Accelerator (Feb - Oct)
An internal transfer, but drastically different work from my previous role, so it felt like starting a new job.
- Evaluated A2A, CrewAI, LangGraph, and other AI frameworks and libraries
- Built a frameworkless agent after internalizing the “agent is just a while loop” insight (ran at half the time and used 1/5th the tokens compared to the framework-dependent approaches)
- Built several agents for simple workflows: Jira automation, automatic data schema, research assistant etc.
- Delivered code to UK and Australia counterparts
- Assisted with packaged service development: Helm charts, Harbor container registry, Kubernetes deployment patterns
- Architected core structure for the legacy VISTA API
- Built a full-stack maritime data visualization system for a prototyped command and control (C2) platform. Backend: FastAPI + PostgreSQL/PostGIS with WebSocket architecture for real-time AIS data streaming.
- Data pipelines: Implemented ingestion for AIS, weather, and mock cargo manifest data
- User research: Planned, organized, and conducted multiple user interviews with warfighters. I gathered feedback and incorporated it into the product roadmap.
Developer Tooling
- PR review summarizer script: extracted comments and used an LLM to summarize and sort by priority.
- Claude Code CI/CD integration: took significant initiative to champion this internally and get it integrated into our pipeline.
Collaboration & Mentorship
- Mentored team intern Mike. Introduced him to the team and our codebase. Consistently pair-programmed and helped with his debugging. Provided a touch point for all things Leidos. Sprinkled in some career advice and book recs :)
- Led regular pair programming sessions with coworkers
- Consistently chose to work in-office specifically to collaborate with local teammates
Plastic Labs (Nov-Dec)
Working for a small AI research lab doing developer relations for Honcho, an AI-native memory and personalization platform.
- Met with builders using the product to understand pain points and gather feedback
- Published integration guides and cookbooks for developers.
- Built a browser extension for granular user activity tracking and assessing the impact of this data in the memory product.
- Dove into memory benchmarks (LoCoMo, LongMemEval, BEAM) and the competitive landscape.
- Daily Twitter publishing. Building a voice for the product and engaging with the AI dev community.
What I learned
- How much I enjoy collaboration. Choosing to work in-office, pair programming, running user interviews. I do better work when I’m not isolated.
- I understand and grow the most when I get my hands dirty.
- What building agents really looks like.
- The complexity and nuance of data pipelines for C2 systems.
- A lot more about twitter and the importance of marketing products well.