Mohamad Abboud
Résumé PDF

00Index Open to full-time roles · Kansas City or remote

Mohamad Abboud

Full-Stack Software Engineer

Five years of end-to-end ownership of client-facing applications, from requirements and architecture through deployment, infrastructure and ongoing support. C#/.NET, Python and PHP, across secure financial transaction platforms and a portfolio of commercial web products, as the sole technical owner and the primary client contact. Since 2026, designing, training and evaluating my own machine-learning systems end to end, without hosted AI APIs.

5
Years end-to-end ownership
400+
Banking institutions served
6
Client applications owned
80%
Processing time removed

LinkedIn GitHub Email Résumé PDF

←→ move sideways scroll works too 1–6 jump to a section

01Experience1 of 3

Independent Projects

Software Engineer

Dates
Jan 2026 – Present
Place
Kansas City, MO
Scope
Self-directed, open source
PythonPyTorchDistilRoBERTaDistilBERTPostgreSQLDockerGitHub ActionsCaddyWebGL2
  1. Trained my own ML models end to end, no hosted AI APIs

    Designed, trained and evaluated my own machine-learning models end to end in PyTorch, with no hosted AI APIs: a two-stage classifier (a DistilRoBERTa detector, then a DistilBERT typer across 13 fallacy types) that flags passages which may lean on a logical fallacy, with a confidence score rather than a verdict (Fallacy Detector).

  2. Invoice pipeline at 98.3% field accuracy vs a 75.1% baseline

    Built a document-intake pipeline that turns messy invoices into validated PostgreSQL records, with a review queue for the documents it is unsure about. Its own evaluation reports 98.3% field accuracy against a 75.1% baseline, covered by 355 tests (Mailman).

  3. AI-chat memory benchmarked on a hand-written 261-fact answer key

    Built portable memory for AI chat: a long conversation becomes a compact brief where every line traces back to the messages it came from, benchmarked against a plain summary at the same token budget on a 261-fact answer key written by hand (Herder).

  4. WebGL2 renderer from scratch, no engine, 620 tests

    Wrote a WebGL2 renderer from scratch, with no engine, for a voxel diorama a camera walks through on a timeline, for recording video, covered by 620 tests (Trail).

  5. Built and run this site: CI/CD, Docker, own domain

    Run this site end to end: requirements written before the code, an allowlist build published by GitHub Actions on every push, and Docker Compose with Caddy for the services that run on a server, on my own domain.

01Experience2 of 3

Tekkii

Full Stack Web Developer

Dates
Jan 2023 – Jan 2026
Place
Kansas City, KS
Scope
Sole technical lead
PythonPHPREST APIsPostgreSQLSQL ServerMySQLEDIPayPalSendGridOneSignalAgile/Scrum
  1. Sole technical lead for six client applications

    Owned a portfolio of six client web applications across logistics, e-commerce, public sector and retail as the sole technical lead, driving each engagement from requirements and client communication through architecture, development, deployment and ongoing support in an Agile/Scrum process.

  2. EDI automation: roughly 80% less processing time, 50% profit increase

    Architected and built an automated EDI capture and transmission system from the ground up for Kansas City Drayage, replacing an entirely manual data-entry workflow, cutting processing time by roughly 80% and contributing to a 50% increase in profit.

  3. Set the backend direction: REST APIs and SQL in Python and PHP

    Set the backend direction across the client portfolio, designing RESTful APIs, service layers and optimized SQL databases (PostgreSQL, SQL Server, MySQL) in Python and PHP, and integrating third-party services including PayPal, SendGrid and OneSignal.

  4. Inherited and stabilized other developers’ applications

    Inherited and stabilized applications built by previous developers, clearing long-standing defects and advising application owners on architecture, technical debt and feature direction.

  5. Rebuilt an e-commerce backend; the growth supported its acquisition

    Led the rebuild of an e-commerce platform’s backend and database layer, driving growth in traffic and revenue that supported the product’s successful acquisition.

  6. Hardware-integrated installation at the College Basketball Experience

    Delivered an interactive, hardware-integrated installation at the College Basketball Experience, personally integrating a custom application with cameras, displays and a teleprompter.

01Experience3 of 3

Allied Engineering Group

Software Engineer

Dates
Jun 2019 – Jun 2021
Place
Beirut, Lebanon
Scope
Financial transaction platforms
C#ASP.NETSQL ServerSWIFTStored ProceduresQuery OptimizationSchema Design
  1. SWIFT platforms for 400+ banks, thousands of secure transactions daily

    Built and maintained SWIFT financial transaction platforms in C#/ASP.NET on SQL Server, processing thousands of secure transactions daily for over 400 banking institutions worldwide, including QNB.

  2. Owned the data layer for high-throughput transaction systems

    Owned the data layer for mission-critical, high-throughput transaction systems, designing schemas and stored procedures and optimizing complex queries to improve application performance and reliability.

  3. Direct technical liaison to banking clients

    Served as the direct technical liaison to banking clients, gathering requirements and delivering customized interfaces and features that measurably increased usability and client satisfaction.

  4. Kept production running by diagnosing issues from SQL and application logs

    Diagnosed and resolved production issues by monitoring SQL and application logs, maintaining uninterrupted operations for banking clients and surfacing risks and blockers ahead of release.

02CapabilitiesPick one to see where it was used

What I work in

01Languages

02Web Development

03Databases

04Architecture & Delivery

05Machine Learning

03ProjectsTwelve things I have built

Projects

04Education

University of Missouri-Kansas City

Master of Science Computer Science

2021 – 2023

Course project

Pneumonia Detection App

A deep learning application for medical image analysis in Python, classifying chest X-rays for signs of pneumonia.

05Contact Open to full-time roles · Kansas City or remote

Let’s build something.

Hand-written HTML, CSS and JavaScript · No framework · No build step · One file