Software Engineer · AI Systems · Quality Engineering
3+ years delivering TypeScript & Python services on AWS & Azure. Shipping fraud detection, RAG pipelines, and scalable APIs that handle millions of requests in production.
I'm a Software Engineer with deep experience building AI-powered products from the ground up — and making sure they work correctly under pressure.
My work spans fraud detection systems, RAG pipelines, NLP extractors, deep learning models, and regulatory analytics platforms — all shipped with observability and quality gates baked in.
I'm drawn to AI products precisely because they break in interesting ways. Model drift, edge-case regressions, latency under load — I've built the pipelines that catch these before users see them. Currently at PwC. MS in CS from UT Arlington.
// 8 real repos · hover cards to flip · click GitHub ↗ to view source
Full-stack healthcare management platform built with PHP (Laravel) and Blade templating. Features patient record management, appointment scheduling, and health dashboards — showcasing end-to-end product engineering across the entire web stack.
CNN-based deep learning model to detect brain tumors from MRI scans.
Convolutional Neural Network trained on MRI scan datasets to classify brain tumors. Practical deep learning applied to medical imaging — demonstrating AI for high-stakes diagnostics.
Predicts diseases from symptoms using 3 ML algorithms with Tkinter GUI.
Ensemble of Decision Tree, Random Forest, and Naive Bayes predicts diseases from user-entered symptoms. Tkinter GUI for interactive use. Demonstrates multi-model ML accuracy.
Content-based image retrieval system trained on CIFAR-10 dataset.
Content-based image retrieval using deep feature extraction on CIFAR-10. Implements similarity search to retrieve visually similar images — foundation for visual search systems.
Neural network implementations and experiments built from scratch.
Hands-on neural network implementations covering architectures, backpropagation, and training strategies — built from scratch to understand core deep learning foundations.
Full e-commerce frontend: shop, cart, product pages, blog in pure HTML/CSS/JS.
Complete multi-page e-commerce site: home, shop, product detail, cart, about, blog, and contact. Responsive design with shopping cart logic in vanilla JavaScript.
Fault-tolerant Two-Phase Commit with failure injection and persistent storage.
Implements 2PC with 1 coordinator and multiple participants. Tests 4 failure scenarios: coordinator crash before prepare, participant decline, partial commit failure, and post-agree crash. Persistent storage for recovery.
SE Management, Maintenance and QA project with full documentation.
Academic SE project covering Software Engineering Management, QA processes, project scheduling with MS Project, and system maintenance documentation. Full SE lifecycle practice.
3 years shipping AI systems in production has taught me exactly where they break and how to surface it fast — from fraud detectors to deep learning medical models.
Evaluating generative outputs for consistency, prompt adherence, and temporal coherence.
Evidently AI + Prometheus surfacing drift before users notice degradation.
Clear reproduction steps, severity classification, and structured docs for cross-functional teams.
Finding exact conditions where models and interfaces behave unexpectedly.
In AI systems, quality means understanding the full behavioral envelope — model behavior at boundaries, degradation under distribution shift, real failure modes for end users.
I bring software engineering rigor: systematic test case design, code-level investigation, and reproducible bug reporting that helps teams fix issues fast.
Study the codebase, trace data flows, identify assumptions. For AI products: understand training distribution and failure modes before writing a single test.
Happy-path tests are table stakes. I design around edge cases, adversarial inputs, latency boundaries, and behaviors product teams often miss.
Clear structured findings with steps to reproduce, expected vs. actual, severity, and investigation direction — not just "it broke."
Work cross-functionally to verify fixes, update coverage, and surface patterns that inform better design decisions.
let's connect
Open to contract and full-time roles in AI engineering, quality testing, and full-stack development. Arlington, TX — open to remote.