Agentic AI & LLM systems
Multi-agent orchestration, tool use, RAG, structured outputs, context design, and human-in-the-loop workflows.
Senior Data Scientist · Ph.D. Computer Science
Dallas, Texas
I build AI systems that hold up in the real world.
I turn ambiguous business problems into reliable, production-grade AI—combining agentic systems, NLP, machine learning platforms, and rigorous evaluation.
AI architecture
Agentic workflows
ML evaluation
Enterprise delivery
0.81weighted F1 for hierarchical NLP classification
12%precision improvement in production modeling
136K+real-world ML tests studied
11.7Kconcurrency mutants evaluated
What I bring
Staff-level thinking, hands-on execution.
I work across the full AI lifecycle—setting technical direction, building the hard parts, and creating the evaluation and operating discipline that keeps systems useful.
Production ML platforms
Data and feature pipelines, experiment tracking, model serving, observability, feedback loops, and automated retraining.
Evaluation & AI quality
Offline evaluation, error analysis, model comparison, robustness testing, guardrails, and measurable release criteria.
Technical leadership
Architecture, technical roadmaps, reusable patterns, cross-functional alignment, and clear decisions under ambiguity.
CORE TOOLKIT
AI & MLLLMs · Transformers · DeBERTaV3 · NLP · Embeddings · RAG · Agent orchestration
PlatformDatabricks · MLflow · PySpark · Azure · AWS · Docker · Linux
EngineeringPython · SQL · C# · Java · APIs · Streamlit · Distributed systems
Selected work
Systems with measurable outcomes.
A selection of enterprise and research work. Confidential details are intentionally kept at the problem-and-impact level.
NLP · LLM · MLOPS
Customer intelligence at production scale
Designed a hierarchical classification and insight platform for high-volume customer feedback, combining GPT-family models, DeBERTaV3, Databricks, and MLflow.
0.81weighted F1
+12%precision
<2sexperience target
AGENTS · RISK INTELLIGENCE
Reliable agentic decision workflows
Designing AI systems for risk signals and complex operational workflows, with explicit attention to orchestration, evaluation, traceability, and human oversight.
ML QUALITY · EMPIRICAL STUDY
How real ML systems are tested
Analyzed 2,525 open-source ML projects and more than 136,000 test cases to identify testing practices, recurring gaps, and opportunities for stronger ML quality.
2,525 projects
136K+ tests
4 ML ecosystems
Experience
AI depth grounded in software engineering.
My path spans enterprise platforms, academic research, and production AI—useful range for making sound architecture decisions and carrying them through delivery.
Request a full résumé-
2026—Present
Senior Data Scientist
GM Financial · GM Protection
Driving applied AI initiatives focused on production ML, intelligent workflows, risk signals, and durable system design. Establishing patterns for orchestration, evaluation, observability, and responsible human oversight.
-
2026
AI Engineer
General Motors R&D · Optimal Inc.
Contributed applied AI engineering in an automotive R&D environment, connecting model development with maintainable software workflows.
-
2025—2026
Data Scientist · Applied AI
DTE Energy
Built production NLP, LLM, RAG, and analytics systems in Databricks. Improved classification precision by 12%, delivered 0.81 weighted F1, and reduced manual data-quality checks by 30%.
-
2021—2026
Doctoral Researcher
Oakland University · Data Science Lab
Advanced software testing for actor concurrency and machine learning systems through large-scale empirical analysis, tool building, and peer-reviewed research.
-
2009—2017
Senior Lead Software Engineer
Poshtiban Niro Co.
Led architecture and modernization of enterprise ERP platforms across HR, inventory, and order management, with a focus on relational data and concurrent processing.
Research
Research that improves system reliability.
My academic work connects empirical evidence with practical tools for testing concurrent and machine learning software.
Featured publicationESEC/FSE 2023
μAkka: Mutation Testing for Actor Concurrency in Akka Using Real-World Bugs
Mohsen Moradi Moghadam, Mehdi Bagherzadeh, Raffi Khatchadourian, and Hamid Bagheri.
A mutation-testing framework grounded in real actor-concurrency bugs, designed to evaluate whether tests can detect failures that matter in industrial-strength Akka systems.
- 186
- real-world bugs analyzed
- 32
- mutation operators designed
- 11,736
- mutants generated
- 10
- real applications evaluated
COMPUTER SCIENCE · 2026
Oakland University
Dissertation: Advancing Software Testing: Mutation Testing in Actor Concurrency and Empirical Insights into Machine Learning Test Practices
COMPUTER SCIENCE · 2020
University of Illinois Springfield
Graduate study in computer science, software engineering, and applied machine learning.
Let’s build something durable
Complex AI problem?
I’m always glad to exchange ideas about applied AI architecture, trustworthy agents, ML quality, and turning research into production systems.