Software engineer · AI systems · platform thinking

Devin
Martinolich

I build reliable platforms, AI-enabled tools, and better developer experiences.

I’m open to conversations about senior and staff engineering work where good systems thinking has to meet useful outcomes.

About

I’m at my best when a problem is important enough to be messy, but concrete enough to finish.

How I approach problems

I start by making the shape of the problem visible: who is involved, what the system needs to do, where the constraints are, and what a useful result would look like. I like alternatives on the table early, then I narrow deliberately. The goal is not to produce the most elaborate design. It is to find the practical path that will still hold up when people have to use and maintain it.

How I work with people

Good technical work is collaborative before it is technical. I want the broader context, I want the people closest to the work in the conversation, and I want disagreement to improve the answer. I bring focus and definition to a group, but I also make room for the information and buy-in that make a solution usable beyond the person who designed it.

How I carry work through

I care about the last mile: integrations, operating details, feedback, failure modes, and the handoff that lets a team keep moving. I prefer visible short-term progress over long stretches of abstract planning. A complete solution is one that connects the right pieces, reaches the people who need it, and can be trusted after the original excitement wears off.

Experience

A career spent connecting product intent to systems that can actually run.

2019—now

Mobile Software Developer III

NYCM Insurance · Java Application Developer II → Mobile Software Developer II → III

Own Android delivery for insured-facing and internal products, support iOS architecture decisions, and partner with CX and product to turn high-priority problems into complete releases.

Built an internal AI springboard used by more than 30 developers, testers, and supervisors; led training on agentic AI tools, prompting, and developer productivity workflows.

Earlier

Android Developer · .NET Developer · Mobile DevOps

FasTraxPOS

Built Android and C#/.NET tooling, modernized legacy retail and mobile software, and supported mobile delivery operations.

Earlier

IT Manager · Business Development · Systems Administration

SC Choice Management

Managed IT operations, infrastructure support, and network/server administration while contributing to business development.

Earlier

IT Manager

Directive

Managed day-to-day IT support and technical operations.

Earlier

Technician

Edison Computers

Performed PC technician work and laptop/mobile hardware repair.

Selected Work

A few systems where the hard part was connecting the pieces.

Developer knowledge baseMake technical knowledge easier to find

Problem

Useful developer knowledge was distributed across systems and difficult to retrieve at the moment of need.

Solution

Designed a knowledge-base system that connected a Java/Liberty service, REST interfaces, MCP, OpenAI APIs, DB2, and vector embeddings into a coherent retrieval experience.

Delivery

Carried the system through integration and Jenkins CI/CD so the answer was not just a prototype, but a path the team could operate.

Agentic ticket triageTurn incoming support work into an organized queue

Problem

Application support tickets arrived with inconsistent context, making classification, routing, and historical discovery expensive.

Solution

Built an orchestration path that extracted historical tickets from HCL Notes Domino, connected to ManageEngine ServiceDesk APIs, and coordinated departments around explicit business rules and statuses.

Delivery

Included LangChain, LangFuse observability, CRM planning, ServiceDesk configuration, and Jenkins CI/CD so the operational workflow was designed alongside the agent.

Internal customer-service assistantGive support teams a faster path to useful answers

Problem

Customer-service knowledge needed to be made usable without asking support teams to manually assemble context for every question.

Solution

Designed an ingestion and retrieval workflow using document extraction, chunking, LangChain, Dify, OpenAI APIs, and a feedback/observability loop.

Delivery

Connected the assistant to a Jenkins delivery path and treated ingestion quality as part of the product rather than an afterthought.