Francis Reiner Maia Martins

Software Engineer | Security & AI Applied to SDLC
MSc Candidate in Software Engineering — Universidade de Brasília (PPCA)
Brasília, Brazil

Research

My current work sits at the intersection of software engineering, security, and applied AI — specifically on how large language models can make secure software development faster, more consistent, and more scalable in regulated environments.

AI-Assisted Security Requirements Engineering & Threat Modeling
Empirical evaluation of a human–AI hybrid workflow for generating security requirements and STRIDE-based threat models, applied to 20 real-world projects in a Brazilian public organization.

• 18.3% reduction in artifact generation time (hybrid vs. manual)
• +13.6% STRIDE threat coverage
• BERTScore semantic precision: 0.73
• Fully on-premise LLM infrastructure with contractual data isolation

Paper submitted to ICSE SEIP 2026 · Under review

Technical Background

Security & DevSecOps Fortify · Tenable · Trivy · Dependency Track · Grype · SonarQube · OWASP · STRIDE · LGPD compliance

CI/CD & Infrastructure Jenkins · Git · Jira Data Center (v10.3.7) · Linux · Docker basics

AI & Automation LLM orchestration (on-premise) · n8n · RooCode · Prompt engineering · Python scripting

Empirical Research Linear Mixed-Effects Models · BERTScore · Fleiss’ Kappa · Paired t-tests

Education

Certifications: CBPP® (Certified Business Process Professional) — ABPMP International · Valid through 2029

Background

Over 18 years of experience across software development, business analysis, and process engineering. Started as a development team lead in Fintech (Cedro Technologies — trading platforms for BM&FBovespa). Moved into requirements engineering, agile team management, and DevSecOps over the following decade.

Currently focused on AI-assisted security workflows and empirical software engineering research as part of my MSc at UnB, supervised by Prof. Elaine Venson (PPCA).