I started programming in 2016 at 10 years old on a Raspberry Pi—flashing OSes, terminal scripting, and breaking systems. By 9th grade, I turned that curiosity into structured engineering with Java—building a solid foundation in OOP, data structures, and system design through high school projects ranging from Android applications to Arabic NLP tools. That "open it up and see how it works" instinct has evolved directly into how I architect production systems today.

Today I'm an ML Engineer and Information Technology student (Dual Studies) at Al-Quds University, recently back from an exchange semester at Mälardalen University in Sweden. I specialize in taking complex ML systems end-to-end—from raw data foundation to production serving. My sweet spot is the retrieval → ranking → recommendation stack: getting the exact right thing in front of the right person.

• 🏗️ Architecting ML Platforms: Interning as an ML Platform Engineer at DevelopOn, designing a scale-ready ML platform monorepo (apps/, packages/, training/, pipelines/) built around a multi-stage candidate-job matching pipeline (hybrid dense/sparse vector retrieval → Learning-to-Rank → Cross-Encoder reranking).

• ⚖️ AI Governance & Engineering Rigor: Leading compliance and architectural governance for the platform under the EU AI Act (Annex III High-Risk) with automated candidate region-gating. Driven by a formal Architecture Decision Record process with 60+ ADRs governing data ownership, model gating, and infrastructure scaling strategies.

• 🎙️ Conversational Agents & Perception Ensembles: Architected questions generation methadogly for an AI interview Agent from job disceiptions and integrated a post-session perception scoring ensemble into an AI Interview Agent. That is including facial expression and emotion classification, gaze tracking for reading detection, and synthetic text/speech detection.

• 🔍 Local, LLM-Free Resume Parser: Engineered a fully local, commercially safe resume-parsing pipeline using layout-aware document extraction, embedding-based section classification, and Named Entity Recognition (NER)—built to handle messy, multi-column PDFs without LLM latency or API costs.

• 👗 Hardening WardrobeGenie: Scaling WardrobeGenie solo from a university project into a production-grade context-aware recommendation engine.

• 📚 MLOps & Agentic AI: Completing DeepLearning.AI's MLOps Specialization (focusing on data validation, concept drift, and human-level performance baselines) while building with LangChain and LangGraph.

I gravitate toward total system ownership over managing a narrow slice. Whether designing a Medallion data lake architecture with automated quality gates, prototyping graph neural network (GNN) matching models, or authoring 80+ ADRs to justify system boundaries, I prefer to go deep on one platform than skim across five.

• Languages & Core: Python (Advanced), Java, Kotlin, Swift, SQL

• AI & ML: PyTorch, Graph Neural Networks (GNNs / PyTorch Geometric), Scikit-learn, Transformers, SBERT, Knowledge Distillation

• NLP, CV & Perception: LangChain, LangGraph, RAG Systems, Facial Expression & Emotion Classification, Gaze Tracking, Speech & Prosody Analysis, Synthetic Content Detection, Object Detection, U-Net Segmentation

• Retrieval & Data Engineering: Hybrid Search (BM25 + Dense Vectors + RRF), Vector Databases (Qdrant, Elasticsearch), Medallion Lakehouse Architecture, Data Quality Validation, Apache Airflow, Polars, Pandas, PostgreSQL, Object Storage (S3/MinIO)

• MLOps & Cloud: FastAPI, Docker, Kubernetes, Databricks, MLflow, Datadog, Terraform, GitHub Actions, GitFlow

• ML Platform Engineer Intern @ DevelopOn (Jun 2026 – Present) — Architected the scale-ready ATS ML monorepo, designed multi-stage hybrid matching pipelines, built EU AI Act compliance gating with 60+ ADRs, and prototyped graph neural network matching models.

• Software Engineer Intern @ Sada Intelligent Solutions (Jul 2025 – Sep 2025) — Developed a spec-driven Android finance app (Kotlin + Firebase) using LLM-assisted development against a Penpot UI/UX design.

• Data Engineer / Python Developer Intern @ ProGineer (for PDF Solutions, CA) (Dec 2024 – Feb 2025) — Built high-performance Polars/Pandas data processing pipelines and Airflow ETL plugins simulating wafer and die-cutting manufacturing processes.

• 3D Printing Service Provider (Oct 2020 – Present) — Operating a local print-on-demand service using Octoprint, Cura, and Fusion 360.

I got into hiking and photography around the same time I started programming, and more recently picked up horse riding—they all draw from the same impulse to notice detail. I also self-host an OpenMediaVault server with Tailscale VPN and Docker containers. When away from code, I'm usually reading, cooking, or watching recorded Stanford lectures (CS183B, CS193P).

📫 Reach me: ridamansour111@gmail.com · LinkedIn