Software engineer · AI product builder

I build AI thatunderstandsthe work.

I'm Bobby Hall Jr, an engineer with 8 years of experience building full-stack products, agent systems, and engineering intelligence that turns complex context into useful outcomes.

Agent systems✦Engineering graphs✦Full-stack products✦AI infrastructure✦Product design✦Agent systems✦Engineering graphs✦Full-stack products✦AI infrastructure✦Product design✦
Selected work · 2026

Systems with a point of view.

Products exploring how people and AI work together, from a daily game about outsmarting an AI to systems that give AI enough context to do real work.

01 / PRODUCTDAILY AI GAME

JailbreakDaily

Trick the AI into telling you the secret.

A new AI to crack every day. Meet the character, pick your moves (Ask, Plant, Impersonate or Decode), pop the locks, and get them to spill. Solo or with a friend. Free, no signup, about 2 minutes.

Game designAI charactersCo-op play
Visit jailbreakdaily.com
02 / PRODUCTAI WORKFORCE

Roster

Give recurring work a real owner.

Roster lets teams hire focused AI employees with a clear role, trusted company context, and permission to use real tools. Work ends in a finished deliverable, with human approval before important actions.

Agent orchestrationHuman-in-the-loopTool integrations
Visit get-roster.com
03 / PRODUCTENGINEERING INTELLIGENCE

Helix Engineering

Software should explain itself.

Helix connects code, decisions, incidents, ownership, and AI-generated changes into a living Engineering Graph, so teams can understand why software exists, what a change affects, and whether it is safe to ship.

Knowledge graphsEvidence-backed AIDeveloper tools
Visit helix-engineering.dev
Engineering approach

Build the product.Earn the trust.

I work across product, interface, data, and infrastructure. The throughline is simple: turn difficult systems into software people can understand and use.

01

Model the system, not just the prompt.

The best AI products understand relationships: the people, decisions, tools, and evidence around the task.

02

Make the work observable.

Useful autonomy needs clear milestones, durable state, and outputs a human can verify, not a mysterious spinner.

03

Keep judgment in the loop.

Systems should move fast on preparation while keeping consequential actions behind explicit human control.

Have a hard system to untangle?

Let's build something that makes sense.

Start a conversationOpen to product engineering, AI systems, and ambitious collaborations.