Enter the password to continue.
The same four things I do feed directly into the workflows B2B teams run every day, from pipeline to lifecycle.
Architecting end-to-end systems that combine models, data, and tooling.
Building autonomous and semi-autonomous agents that take real actions.
Shipping AI-native products people actually use, day to day.
Working fluently across providers and picking the right model for the job.
Draft and enrich pipeline records automatically.
Flag risk and surface what needs attention.
Generate, target, and iterate on campaigns.
Move customers through each stage automatically.
Turn raw activity into readable reporting.
Provider-agnostic by default, I reach for the model and tooling that fit the constraint: cost, latency, and reliability.
Reasoning & agents
Multimodal & long context
General-purpose workhorse
Low-latency inference
Four production systems across agents, apps, and automation. The stack cycles through each, hover to pause, or pick one.
One prompt in, a full set of platform-ready creatives out. Copy, calls-to-action, and every ad size, built for small and local businesses.
Type anything and play it back. This runs entirely in your browser using the Web Speech API.
Provider-agnostic, cost- and latency-aware. The best model is the one that fits the constraint.
Beyond chat, tool use, integrations, and reliability in real B2B workflows.
Evaluation, guardrails, and the unglamorous work that turns demos into systems people trust.
