Where I've worked.

LLM Routing System (Internal AI Tool)

• Designed and implemented an intelligent LLM routing layer that analyzes incoming prompts and tasks using a local LLM running on-device.
• Evaluates prompt complexity and task type, then dynamically routes each request to either a local model or a cloud-based provider.
• Built the routing decision entirely on local inference, with no external API calls in the decision path.
• Demonstrated measurable potential cost reduction for teams with mixed AI workloads while preserving quality for tasks that require frontier models.

Vehicle Software Infrastructure

• Built and maintained CI/CD pipelines for containerized application deployments on embedded Linux platforms across multiple hardware architectures.
• Developed and deployed containerized backend services on custom embedded hardware targets.
• Implemented an Over-the-Air (OTA) update mechanism in Rust, self-taught on the job, with automated version checks, fallback logic, health monitoring, and service orchestration.
• Automated build, tagging, deployment, and update workflows across development and target environments.

Edge AI Computer Vision Prototype

• Developed an edge-based computer vision prototype for real-time dashcam footage processing.
• Built the perception pipeline using GStreamer with YOLO for object detection (vehicles, pedestrians, environmental elements).
• Entire pipeline ran locally on embedded edge hardware — low latency, no cloud dependency.
• Implemented preprocessing, inference, post-processing (NMS, filtering), and output generation stages.
• Produced two outputs: processed video stream with bounding-box overlays and structured per-frame detection metadata.
• Metadata drove simplified visual feedback (ambient light signals) representing nearby objects or hazards.
• Concept explored real-time visual perception supporting driver awareness in low-visibility conditions.

Confidential AI and Technical Evaluation Project

• Contributed to an early-stage confidential internal project where technical direction was still being shaped.
• Investigated and compared tools, technologies, hardware and software options, and AI-related approaches.
• Tested selected options, documented observations, and prepared internal reports and summaries.
• Specifics are under NDA.

Recognition

• Worked as a full team member in a cross-functional Scrum team — sprint planning, code reviews, and architectural discussions.
• Recognized by my supervisor for high initiative, structured problem-solving, and clear technical communication.
• Academic supervisor Dr. Ahmad Barghash ranked me in the top 5% of graduation-project students he has supervised.

Contributions

• Developed RESTful APIs using ASP.NET Core for mobile and web applications.
• Designed relational data models and implemented CRUD using Entity Framework Core and MySQL.
• Optimized API performance through query tuning, caching, and structured backend refactoring.
• Worked in an agile team with Git-based version control and code review workflows.
• Collaborated with frontend and product teams to integrate backend systems with user-facing features.