Products / Production systems / Applied AI

Engineering work

Hands-on software and reliability engineering across Microsoft cloud platforms, a deployed product and open-source systems work.

Download CV View GitHub Email Jason

More than a decade at Microsoft Software through production operations Based near Seattle

Jason Doyle / Monroe, Washington

01 / Selected products

Built far enough to encounter the difficult parts.

These projects were built independently outside my Microsoft role. They cover customer workflows, access boundaries, recovery behaviour and production operation.

01 / Deployed product

BarnPage

Visit BarnPage

BarnPage is a private workspace for riding barns. I designed and built the product from its data model and role boundaries through deployment and operation.

Product

Scheduling, barn membership, horse and rider records, private lesson media, progress feedback and trainer-reviewed AI workflows.

Engineering

TypeScript, Next.js, PostgreSQL, Prisma and Azure. A separate worker handles video processing, while scheduled workflows handle operational tasks.

Boundaries

Barn-scoped permissions, guardian links, consent controls and private media access are part of the product model rather than optional policy text.

AI use

AI output remains draft material for trainers. Riders and parents see feedback only after a qualified trainer reviews and publishes it.

02 / Open-source systems work

Agentic Data Kernel

View repository

Agentic Data Kernel is a TypeScript persistence and workflow layer for agents that need durable state, evidence lineage and controlled external effects.

Problem

Long-running agents need to preserve what was known, why an action was authorised and whether a timed-out external effect actually happened.

Implementation

SQLite supports local use and PostgreSQL supports the production profile. TypeScript, HTTP and MCP expose the same typed operation model.

Recovery evidence

The flagship scenario restarts runtime state, reconciles an uncertain rollback and verifies recovery without delivering the action twice.

Tradeoffs

The repository publishes raw comparison evidence and states where ordinary PostgreSQL matches the kernel or remains the simpler choice.

Read the recovery scenario View the package

02 / Professional impact

Production work with measurable consequences.

Public descriptions are limited to evidence that can be shared without disclosing proprietary systems.

Incident response

Approximately 40% lower time-to-resolution

Contributed through structured incident command, telemetry-led diagnosis and standardised follow-through across a multi-service portfolio.

Telemetry platforms

Shared evidence across cloud services

Architected analytics foundations using Azure Data Explorer, Azure Synapse and Microsoft Fabric.

Customer reliability

Signals people can use

Defined an approach for turning internal service indicators into clearer customer-facing reliability information.

03 / Experience

Software, data and reliability across Microsoft.

Roles moved from implementation and automation into service ownership, incident leadership and cross-product reliability.

Senior Customer Experience Engineer

Technical Lead, Reliability and Observability

Cross-product reliability, telemetry platforms, service-health signals, customer incidents and applied AI analysis across Azure.

March 2025 - Present

Senior Site Reliability Engineer

Azure Data

Reliability strategy for production data services used by millions, including high-severity incident response and operational improvement.

September 2023 - March 2025

Senior Site Reliability Engineer

Localization as a Service

Reliability, performance, automation and operational efficiency for high-availability platforms supporting global Microsoft workflows.

September 2019 - September 2023

Software Engineer

Microsoft Ireland

Data processing, telemetry, localisation systems, validation workflows and internal engineering tools for Microsoft Office.

May 2014 - September 2019

Download CV View machine-readable resume

04 / Technical scope

Depth in reliability, with range across the stack.

Skills are listed where they are supported by professional or shipped project experience.

Software and systems

C#/.NET, TypeScript, SQL/T-SQL, APIs, distributed systems, CI/CD and automation.

Production reliability

SRE, high availability, incident command, deep diagnosis, SLI/SLO design and operational readiness.

Data and observability

Azure Data Explorer, SQL Server, PostgreSQL, Microsoft Fabric, Azure Synapse and telemetry platforms.

Applied AI

Agent workflows, MCP, durable state, retrieval, lineage, evaluation and controlled automation.

05 / Technical writing

Writing that exposes the operating model.

The whitepapers separate evidence, assumptions and unknowns, then provide practical templates for implementation and review.

MCP Tool Readiness When the Endpoint Is Up but the Product Is Failing Observability Is a Decision System View all whitepapers

06 / Contact

Talk to me about difficult engineering work.

For hiring conversations, include the role or problem area and a link to the relevant description.

Start a hiring conversation [email protected]