What I build
Secure web applications, cloud-native services, ML APIs, and automation pipelines where performance and operational reliability are non-negotiable.
Software Engineer | AI/ML | Cybersecurity
I build secure, production-ready systems with measurable outcomes.
I am a full-time Ph.D. candidate in Computer Engineering at Texas A&M and a part-time software engineer. My work combines cloud architecture, secure backend engineering, and applied machine learning for real operational problems.
About
I work at the intersection of software engineering, AI, and cybersecurity. I prefer shipping systems that are understandable, testable, and secure by default rather than over-designed demos.
How I approach engineering
- Define success in measurable terms before implementation.
- Design for least privilege, abuse prevention, and observability from day one.
- Prioritize maintainable architecture and predictable deployment workflows.
Experience
Roles where I have shipped production systems and applied research in real environments.
Software Engineer, Avika Billing Solutions LLC
- Architected and deployed a HIPAA-aligned SaaS medical billing platform using Flask, AWS Elastic Beanstalk, RDS, and SES for 100+ active users.
- Designed ACH and credit card payment infrastructure (NACHA generation, webhooks, monitoring), reducing processing time by 70%.
- Built deployment and infrastructure workflows including SSL/TLS, domain management, CI/CD, and database migrations.
- Configured AWS WAF and rate limiting policies to prevent abuse while supporting a 99% uptime target.
Graduate Researcher, PRISE Project (ECEN), Texas A&M University
- Designed and trained a PyTorch model for malicious TCP traffic detection with 95% accuracy.
- Processed 100K+ network packets with Pandas and NumPy to identify anomalies and attack patterns.
- Deployed models as Flask REST APIs for real-time inference in simulated production settings.
Graduate Teaching Assistant, ETID, Texas A&M University
- Led networking laboratories with Cisco hardware and practical routing and switching exercises.
- Built topologies demonstrating OSPF, BGP, RIP, VLANs, trunking, DHCP, and STP.
- Mentored students through troubleshooting workflows on Linux and Windows terminal environments.
Education
College Station, TX | GPA: 3.8
College Station, TX | Graduated Dec 2024 | GPA: 3.7 | Thesis: Download
College Station, TX | Graduated May 2023 | GPA: 3.5
Relevant Coursework
AI - Deep Learning - Reinforcement Learning - Data Mining - Intelligent Agents - Data Analysis & Experimental Methods - Data Analytics for Cybersecurity - Embedded Systems Software Development - Embedded Systems Intelligent Design - Project Management - Advanced Network Systems & Security - Industrial IoT.
Featured Projects
Each project summary highlights context, implementation scope, engineering decisions, and results so technical reviewers can evaluate depth quickly.
Reinforcement Learning Firewall
View RepositoryProblem / Context
Static firewall rules struggle with evolving traffic behavior and can create costly false positives or missed attacks.
What I built
Created a custom Gymnasium environment and trained Q-Learning, SARSA, and DQN agents to make sequential allow/deny decisions.
Engineering decisions
Designed a reward function that balances detection performance against operational cost, then compared learning approaches against static baselines.
Impact / Result
Demonstrated stronger policy adaptation than fixed-rule baselines in simulation.
Stack: Python, Gymnasium, reinforcement learning algorithms.
ASIC: AI for Satellite Image Classification
View ProjectProblem / Context
Satellite segmentation tasks require robust classification across varied terrain and image quality constraints.
What I built
Fine-tuned a semantic segmentation pipeline on 800+ satellite images covering seven land-cover classes and exposed it through a Flask web app.
Engineering decisions
Focused on model and preprocessing choices that supported practical user-driven image analysis through a web interface.
Impact / Result
Reached 70%+ pixel-level segmentation accuracy on real-world imagery.
Stack: Python, Flask, computer vision segmentation workflows.
Dynamic Honeypot Game
View RepositoryProblem / Context
Defenders need adaptive strategies that anticipate attacker behavior in constrained network environments.
What I built
Modeled attacker and defender interactions as a two-player zero-sum game and used multiplicative weights update to approximate Nash equilibrium behavior.
Engineering decisions
Simulated neighborhood detection and strategy shifts in a grid network environment to evaluate tactical tradeoffs.
Impact / Result
Produced a simulation framework for comparing dynamic deception strategies under adversarial conditions.
Stack: Python, game-theoretic modeling, simulation tooling.
What I Work With
Backend and platforms
Python, Flask, FastAPI, JavaScript, SQL (MySQL/PostgreSQL), API design, authentication and role-based access patterns.
Machine learning and data
PyTorch, TensorFlow, Keras, scikit-learn, NumPy, Pandas, and model-to-API deployment for applied ML workflows.
Cloud and operations
AWS (Elastic Beanstalk, EC2, RDS, S3, SES, CloudWatch), CI/CD pipelines, TLS setup, migration workflows, and production monitoring.
Security and systems
AWS WAF, rate limiting, network analysis, Linux-based tooling, Wireshark, and cybersecurity-focused experimentation.
Languages
Portuguese (Native), English (Fluent), Spanish (Advanced).
Contact
For full-time roles, research collaborations, and engineering engagements.
Reach me directly
The easiest way to start is email with a short brief: your team context, the technical problem, and desired outcome.
What I can help with
- Designing and hardening cloud-based applications.
- Shipping ML features as reliable services, not isolated notebooks.
- Improving security posture without slowing delivery.