Alex Haas / A bit of background

I tend to follow
the problem.

I’m a software and systems engineer in New York. I design APIs and data models, build applications and automation, and diagnose production failures across code, networks, and infrastructure.

Download résumé

Read a few technical investigations ↓

My path has included military service, construction, technical support, and network engineering. Software brought those interests together: people, practical constraints, and the machinery underneath.

I’ve been the only technical support person for a company, worked on hotel networks, and built tools used in a large healthcare environment. Outside that work, I’m developing Socraites, volunteering on DDRC Workbook, and building Filament Friends.

Inside the work

A few problems
I’ve worked through.

I don’t assume a problem belongs to just one layer. These are examples of following it through the code, the network, and the behavior people actually see.

Windows internals / Application debugging

The application was choosing the wrong network interface.

What broke: Adding a second network interface for backups caused a vendor application to fail.
How I traced it: I reverse-engineered the application’s interface-selection behavior and traced it to a Windows call that selected the last-installed NIC—even when it belonged to the wrong network.
What changed: Reconfiguring the interfaces made the application select the production network again. The failure sat at the boundary between application behavior and infrastructure.

Wireshark / Active Directory / Network diagnosis

The slow application was waiting on authorization traffic.

What broke: After an Active Directory migration, applications became slow. One page checked authorization for many user and service accounts, amplifying the delay.
How I traced it: Packet captures showed long waits while requests went to an AWS-hosted domain controller instead of the local on-premises controller. I identified the path and brought the evidence to the domain-controller team.
What changed: The team repointed the application servers to local controllers. The OCR web application lost more than two minutes of delay, and associated MoveIt errors cleared.

API design / FastAPI / Performance testing

A new API contract had to work against a mainframe.

What broke: A call-center application needed access to member, provider, eligibility, claims, and treatment-authorization information held on an IBM mainframe.
How I traced it: I designed 14 endpoints, implemented a Linux/FastAPI prototype, and wrote Python performance tests to establish a baseline. Measurements on the mainframe implementation could then be compared against that baseline.
What changed: The production APIs were implemented on IBM z/OS and support an application used by more than 500 agents. My contribution was the API design, prototype, and performance work; the production implementation moved to the mainframe.

Designing systems, too

The decisions behind
the personal projects.

DDRC: keeping reviewed state consistent.

Housing, seating, and work assignments interact. Staff configure a plan, preview it, and explicitly apply it. Publishing reads that applied state; it doesn’t quietly rerun the assignment algorithms.

Check-in commits locally first, then uses a transactional outbox for upstream synchronization. Failed delivery stays visible and can be retried without losing the local arrival record.

DDRC project notes ↗

Socraites: checking what a model claims.

The experimental reasoning pipeline connects proposed interpretations to source records. Code validates references and timing; comparison episodes help challenge a proposed pattern.

A structurally valid answer can still be a bad interpretation. I keep the observations, model proposals, and user feedback separate rather than treating a model’s confidence as evidence.

Socraites project notes ↗

Work I’ve done

Code, people,
and everything
in between.

Senior Professional Application Designer

Gainwell Technologies

Built an OCR test-document tool that reduced preparation from about 15 minutes to about one minute per document, used by a team of six.
Created a document-retrieval tool used by more than ten colleagues, replacing a roughly three-minute web workflow with downloads in under five seconds.
Designed 14 APIs for an application supporting 500+ call-center agents. Built and performance-tested the Linux/FastAPI prototype; the production implementation moved to IBM z/OS.
Helped architect backup and recovery across two data centers and managed the installation and release.
Investigated difficult application and network failures using packet captures, code analysis, and detailed infrastructure diagrams. Built monitoring and remediation scripts for production workflows.

Systems & network engineering / Managed services

TAUTO

Installed and supported switches, routers, cabling, and camera-recording systems across hotel properties.
Worked independently on site during outages, tracing routing and connectivity problems and restoring service.
Served as the Bitdefender subject-matter resource across the hotels the team supported.

Head of Technical Support

Watermark Designs

Was the sole technical support person for more than 50 employees, covering applications, networks, hardware, and onboarding.
Used Python to reshape documents and CSV data for QuickBooks workflows, moving processing from the following week to the next day.
Traced stalled processing through output data so work could resume at the right point.

See my current projects ↗

What I work with

Across the stack.
Into the details.

Applications & integrations

Python, FastAPI, TypeScript, React, Next.js
REST API design, prototypes, and performance testing
Internal tools, desktop utilities, and connections to existing systems

Data & workflows

PostgreSQL, SQL, data modeling, and migrations
Graph relationships with Neo4j; Redis and RabbitMQ
Imports, synchronization, document generation, and image processing

Systems & troubleshooting

Linux, Windows Server, Active Directory, AWS
Routing, packet captures, and application/network diagnosis
Backup and recovery, monitoring, and production incident work

Automation

PowerShell, Bash, Python, and .NET tools
Document processing and API-driven workflows
Checks that inspect actual results, with retries and useful failure information

Applied AI

OpenAI APIs and local model experimentation
AI-assisted development, with review and validation
Image analysis and experimental, source-grounded reasoning in personal projects

Making the system understandable

Detailed architecture and network diagrams
Translating operational problems into working tools
Explaining technical behavior to colleagues with different backgrounds

Study & credentials

Some of the
formal foundations.

Computer science, with other interests.

B.A. in Computer Science, cum laude, Hunter College. Earlier coursework in psychology at the University of Washington.

I’m still drawn to questions about how people think. Socraites is one place where that interest meets the software work.

Learning by making.

My home lab, personal projects, and artwork give me room to try ideas, get things wrong, and understand them better.

The art side of things ↗

Have something in mind?

I’m open to projects, collaborations, and good questions.

mralexhaas@gmail.com

I tend to followthe problem.