About Nevyn Duarte

I'm Nevyn Duarte, an AI/ML engineer and technical leader building production AI systems. I combine applied machine learning, backend engineering, and client delivery to turn ambiguous requirements and complex data into useful products.

Professional Background

I'm Co-Founder, CTO & Lead AI Engineering Consultant at Bridges AI Consulting, where I work directly with clients to scope problems, design systems, and deliver AI applications. My work includes content understanding and generation, decision support, and workflow automation, with human review and data access controls built into delivery.

Prior to Bridges AI, I worked at M Science (Jefferies) as a Quantitative Equity Research Associate, where I developed predictive equity models using PySpark and SQL on Databricks. I analyzed millions of transactions and job postings data points, produced data-driven research reports that accelerated report cycles from quarterly to monthly, and automated data extraction workflows that improved team efficiency by 20%.

My experience also includes roles at Citco Fund Services where I supported risk and valuation models for hedge fund portfolios exceeding $10B in assets under management, and at AMD where I designed adaptive machine learning algorithms for yield analysis across 100k+ wafer samples.

Education

I earned my Master of Science in Data Science from the University of Colorado Boulder in August 2026 with a final GPA of 3.88, focusing on deep learning, parallel computing, and advanced statistical modeling. My coursework included probabilistic modeling, unsupervised machine learning, natural language processing, and generative AI.

I hold a Bachelor of Science in Mathematics from the University of Texas at Austin, where I was recognized with the UT CNS Award for Excellence in Computer Science for my research contributions to the Building-Wide Intelligence (BWI) robotics project.

How I Approach AI Delivery

I start with the workflow and the people who use it: what decision needs to improve, what a useful output looks like, and where errors matter. That framing guides the choice between conventional software, a model, or an AI agent—and the tradeoffs between quality, latency, cost, and operational complexity.

My engineering priorities are clear evaluation criteria, reviewable outputs, reliable integrations, and visibility into failures. I treat deployment, monitoring, and iteration as part of delivering a useful system, alongside the model itself.

Research & Interests

My research interests center on the intersection of machine learning theory and practical applications. I'm particularly drawn to problems in predictive modeling for financial markets, autonomous systems, and natural language understanding. I actively study academic papers and textbooks to inform how I design models, evaluate risk, and build production-ready systems.

At UT Austin, I contributed to robotics research developing trajectory-based person-following systems using DeepSORT tracking and Google's Triplet Loss function, programming robot navigation in Python, C++, and ROS.

Technical Expertise

My backend toolkit includes Python, FastAPI, Go, REST APIs, PostgreSQL, and Redis. For cloud infrastructure and delivery, I work with AWS, Docker, Terraform, and serverless services and queues. My ML and data work uses PyTorch, scikit-learn, XGBoost, HuggingFace, PySpark, Databricks, and SQL, supported by a foundation in statistics and quantitative research.

Let's Connect

I'm exploring applied AI, forward-deployed engineering, and AI/ML platform roles in Austin, including backend and solutions engineering on AI teams. I also welcome conversations about financial ML and quantitative research. If your team is turning a complex workflow into a production system, I'd be glad to connect.