AI Systems Architect

I design and develop independent AI engineering projects focused on LLM governance, agent state and memory management, deterministic decision systems, and applied machine learning.

I build Python libraries, FastAPI services, reference implementations, interactive demonstrations, data-processing pipelines, API interfaces, and testable backend logic. My work explores RAG, model orchestration, evidence traceability, human review, observability, and policy-driven evaluation workflows.

I maintain a portfolio of reusable libraries, technical reference implementations, and hosted demonstrations using technologies such as Python, FastAPI, PostgreSQL, TypeScript, Docker, and cloud-oriented infrastructure.

Contributed to the modernization of CGM MUSE, a large hospital logistics platform with more than 150 modules and 300 workflows. Supported legacy refactoring, Angular development, and reusable software architecture.

Designed and implemented REST APIs, business logic, approval workflows, access controls, and integrations for hospital inventory and operational processes. Developed ETL pipelines, data transformations, reporting datasets, and SQL optimizations for large-scale healthcare logistics data.

Developed and evaluated ML and LLM prototypes, including RAG workflows, training datasets, medication inventory reporting, and rule-based email and in-application alerts. Produced testing and technical documentation for these capabilities.

Independently architected and developed RT-Dispatch, a major GTMS platform module, with ownership spanning UI/UX design, implementation, integration, and delivery.

Built dispatch interfaces, backend APIs, operational workflows, and business logic using Next.js and NestJS. Developed machine-learning-based dispatch forecasting using five years of historical data containing approximately 10 million dispatch records across 2,000 vehicles and 5,000 drivers.

Deployment and testing teams reported an approximately 30% improvement in time-to-dispatch prediction accuracy after implementation at Toronto Pearson International Airport.

Worked across 12 AI and industrial IoT modules supporting operational safety, fleet monitoring, and industrial efficiency applications.

Contributed to machine-learning components for forklift tracking, safety monitoring, anomaly detection, and predictive operational analytics. Worked with telemetry, location, sensor, and operational-event data for real-time monitoring and ML workflows.

Supported integration between AI software, IoT devices, and industrial systems while considering data quality, latency, and operational reliability.

Developed and maintained high-volume banking and payment infrastructure, including systems and services delivered to Bank Maskan. Supported Shaparak payment-switch development and maintenance with a focus on transaction integrity and payment-network compliance.

Led the technical resolution of a critical Shaparak incident involving more than $3 million in unresolved transactions within the required regulatory window. Led development of an early RESTful Internet Payment Gateway in Iran using Node.js and Angular.

Designed ATM monitoring functionality and contributed to banking loyalty and transaction-management systems. Participated in banking-switch database modernization, including migration from relational to NoSQL technologies for scalability.

Built backend systems for succession-planning software for the Oil Industry Management Development Center and performance-evaluation software for the Housing Foundation.

Integrated authentication using CAS, OpenID Connect, and custom SSO providers including FANAP, Maskan Bank, Melli Bank, and MyTehran. Integrated bank payment and SMS services.

Investigated and improved SQL Server performance bottlenecks for large organizations, including the Iranian Social Security Organization and the Agriculture Organization.

Bachelor of Science degree in Information Technology Engineering.