Mohammad Fares

MSc Data Science graduate working as an AI Engineer at Puffersoft, building and deploying production AI systems for enterprise clients. Current work includes a bilingual (Arabic/English) RAG document Q&A system over Milvus, agentic tool calling workflows orchestrated through n8n with MCP connectors, automated LLM evaluation pipelines with hallucination detection, and an LLM inference stack on Kubernetes with a single GPU. Dissertation research evaluated LLM faithfulness across 40 medical triage and financial loan cases, achieving faithfulness scores up to 4.85/5.

Comfortable across the full AI stack, from Python ML pipelines and cloud deployment (AWS, Azure, Docker) to prompt engineering, LLM evaluation, and client facing delivery.

AI Engineer - Puffersoft

(2026-02)

• Built a bilingual (Arabic/English) RAG document Q&A system over Milvus, letting enterprise clients query internal PDF knowledge bases in natural language and replacing hours of manual review per week.
• Ran a frontier LLM evaluation (Qwen 3.7-Max vs GPT-5, 50 questions) with an automated LLM as judge pipeline, hallucination detection, and an interactive dashboard, directly informing model selection for production.
• Integrated Claude with n8n workflows via MCP connectors, enabling tool-calling and agentic multi-step orchestration across enterprise client environments.
• Designed and deployed an end to end CV screening workflow in n8n on Kubernetes that parses submitted PDFs, scores candidates against role profiles with a weighted rubric, stores results in PostgreSQL, and emails shortlists to hiring managers.
• Built additional agentic n8n workflows including an AWS AI news aggregator and automated email pipelines, reducing manual processing time across 3 client environments.
• Deployed an LLM inference stack on Kubernetes with a single GPU and maintained CI/CD pipelines on AWS and Docker, enabling cost efficient, zero downtime production serving across 2 client environments.
• Managed Milvus vector DB infrastructure on Kubernetes (Attu), handling collection design, role permissions, and multi-tenant troubleshooting.
• Prepared and delivered client-facing presentations on Puffersoft's n8n AI Lab platform covering value proposition, use cases, and FAQ.

Software Engineer Intern - Blue Ray for Web Solutions

(2024-05 - 2024-09)

• Developed and delivered full-stack web features (frontend and backend) across 3 client projects within Agile sprint cycles, supporting timely product releases.
• Built responsive UI components that improved mobile usability, contributing to measurable gains in user engagement across client applications.

MSc - Data Science - Liverpool Hope University (2025-02 - 2026-02)

BSc - Software Engineering - Jordan University of Science and Technology (2020-10 - 2024-09)