Here is your updated "Honest Resume," which incorporates your latest roles at Builder.ai and AI Andromeda Ltd., and provides a comprehensive, behind-the-scenes context for a career consultant. This document is designed to be a private tool to help you strategize your job search, tailor your official resume, and prepare for interviews by having a clear and honest picture of your journey. ### Personal Details Name: Arash Behmand Email: arash.behmand@gmail.com Phone: (+44) 759 3359 776 Location: London, UK GitHub: [https://github.com/arashbehmand](https://github.com/arashbehmand) LinkedIn: [https://www.linkedin.com/in/arashbehmand/](https://www.linkedin.com/in/arashbehmand/) --- ### Summary A highly motivated Senior Machine Learning Scientist with nearly a decade of experience leading and contributing to impactful, data-driven projects from ideation to deployment. My expertise spans machine learning, data science, and data engineering, with a strong focus on deploying ML models and leveraging Large Language Models (LLMs) for real-world applications. I am passionate about solving complex challenges using AI and data, having successfully improved fraud detection, optimized pricing strategies, and enhanced customer engagement through innovative machine learning solutions. Proficient in Python, PySpark, SQL, AWS, and cutting-edge AI frameworks like LangChain, I am adept at building scalable data infrastructures and deploying robust ML models. I am continuously learning and innovating, eager to contribute to cutting-edge AI advancements and thrive in a collaborative, forward-thinking environment. I am actively working on improving my communication skills to better translate complex technical concepts into actionable insights. --- ### Experience #### Builder.ai Role: AI / ML Scientist Dates: January 2025 - May 2025 Location: London, UK * Pioneered AI-driven automation of feature documentation at scale for over 500 projects, designing end-to-end Python/LLM systems that reduced manual documentation from 10+ hours to 1 hour per project. * Enabled ingestion of complex design whiteboards and multi-source data, significantly enhancing requirement coverage, product understanding, and team efficiency. * Automated end-to-end feature note generation via a Python/LLM system (LangChain), cutting manual effort from 10+ hours to a 1-hour review per project across more than 500 projects (30+/month). * Engineered advanced LLM pipelines, including an 'Aggregate of Experts' pattern, achieving 88% F1 accuracy and 85% requirement coverage by synthesizing diverse technical inputs. * Architected a multimodal AI system interpreting complex SVG whiteboards (70+ components) using LLM vision models, producing structured markdown blueprints and Mermaid.js graphs. * Built a scalable Python application for automated feature note generation, featuring robust Pydantic data modeling, centralized configuration, and Langfuse integration for LLM monitoring. * Implemented advanced LLM-driven analytical capabilities, including fact verification against technical documents and pruning feature notes for accuracy, providing deeper product insights. * Developed NLP techniques for clustering customization requests, identifying common feature improvement themes, and analyzing product manager note-taking templates for process optimization. * *Context: Despite the company's sudden collapse into bankruptcy due to alleged financial mismanagement, my work on feature note generation became critical, with the entire 8-person NLP team focusing on the system I initiated.* #### AI Andromeda Ltd. Role: Co-Founder & Machine Learning Engineer Dates: April 2024 - December 2024 Location: London, UK