Services

I take technically demanding work from research and architecture through implementation, testing, and production support. An engagement can focus on one difficult part of an existing system or the delivery of a new product.

Product engineering

I design and build web applications, APIs, and backend services. This includes frontend architecture, service boundaries, third-party integrations, payments, subscriptions, administration tools, and modernization of existing systems.

Approach: I start by mapping the users, critical journeys, existing constraints, and integrations. Then I define service boundaries and a delivery sequence, build a usable vertical slice, and check accessibility, behavior, and maintainability before each release.

First deliverable: A working vertical slice and a clear plan for the next release. See the dg-builder editor for an in-progress product example.

AI orchestration

I connect local and hosted models to useful product workflows. I design model routing, retrieval-augmented context, tool-based flows, and human review points, with attention to evaluation, cost, and operational control.

Approach: I identify the decision a model should support and the data it may use. Small representative cases guide model and tool selection; I design permissions, human approval, failure paths, and observable cost into the workflow before expanding it.

First deliverable: A small evaluated workflow with an explicit human review point. See dose-ai-gateway for the local orchestration direction.

Data and performance engineering

I improve how systems store, synchronize, and serve data. Relevant work includes PostgreSQL schema and query tuning, indexing, event-driven services, caching, and offline-capable browser data flows.

Approach: I measure the slow or unreliable path first, inspect query plans and data ownership, then make targeted changes to schemas, indexes, caching, or synchronization. I compare results against the baseline and keep migration and rollback steps explicit.

First deliverable: A measured baseline, the likely bottleneck, and a prioritized change plan. An anonymized performance case shows the approach in production.

Interactive web experiences

I build browser-based 3D and extended-reality experiences with content that editors can manage. My previous work includes Three.js, Babylon.js, WebXR, and custom content-management integrations.

Approach: I prototype the interaction and visual language early, then choose the rendering method for the device and content. I test motion, loading, input, accessibility, and reduced-motion alternatives together so the experience remains usable beyond a desktop demo.

First deliverable: A device-tested interaction prototype with a reduced-motion alternative. The isometric editor shows this kind of browser-based 3D work.

Delivery and reliability

I make development and deployment more repeatable with CI/CD, Linux, Docker, Nginx, monitoring, and clear handoff practices. Scope can include diagnosing production bottlenecks and reducing operational friction.

Approach: I document how a change moves from a developer machine to production, automate repeatable steps, and add checks where failures are costly. I include health signals, a recovery path, and a handoff that lets the operating team understand the system.

First deliverable: A deployment map with its most fragile steps and a practical recovery path. The business-continuity case shows delivery under an operational constraint.

Preliminary Incident Response report

When an incident needs an initial technical assessment, I can review available logs and system evidence, build a preliminary timeline, record likely impact and open questions, and recommend immediate next steps. The report is an initial engineering assessment; deeper forensic investigation or legal conclusions require a specialist engagement.

Approach: I agree on the question and evidence available, preserve a clear record of sources and timestamps, and separate observed facts from hypotheses. I deliver a concise timeline, impact assessment, confidence and gaps, and prioritized containment or follow-up actions.

First deliverable: An evidence inventory, preliminary timeline, and immediate questions to resolve. Start a conversation to define the scope before any assessment.