Principal Software Engineer - Breeze Assistant

HubSpot is an AI-powered customer platform that enables businesses to grow faster by connecting marketing, sales, and service. They are seeking a Principal Software Engineer to lead the Breeze Assistant product line, focusing on developing AI-driven solutions that enhance user interactions and streamline workflows.

Responsibilities:

• Architectural Leadership for AI Actions: Define the patterns for how AI agents safely and predictably interact with HubSpot’s CRM and third-party ecosystems (Slack, Google Workspace, Microsoft 365)
• Rapid GenAI Experimentation: Act as a "zero-to-one" builder. You stay current with the latest trends in LLMs, RAG (Retrieval-Augmented Generation), and Agentic frameworks (like LangGraph or AutoGen) to rapidly prototype and validate new "System of Action" capabilities
• Hands-on, High-Leverage Contributor: You ship production code. You will lead the development of high-impact initiatives, moving from a Jupyter notebook prototype to a petabyte-scale production system with ease
• System Reliability & Trust: Ensure that GenAI outputs are not just "cool," but trustworthy. You will drive the strategy around LLM evaluation, observability, and guardrails to ensure accuracy and data privacy
• Technical Mentorship & Multiplier: You will elevate the AI fluency of the entire engineering org. You’ll mentor senior engineers on prompt engineering, model selection, and the nuances of building non-deterministic systems
• Cross-Functional Influence: Partner with Product and AI Research teams to scope opportunities where GenAI can solve real-world GTM friction, ensuring our technical bets align with HubSpot’s broader AI strategy
• Rapid Domain Fluency: The AI landscape changes weekly. You must be able to dive into a new paper, model, or framework today and explain its strategic value to the team tomorrow

Requirements:

• Architectural Leadership for AI Actions: Define the patterns for how AI agents safely and predictably interact with HubSpot's CRM and third-party ecosystems (Slack, Google Workspace, Microsoft 365)
• Rapid GenAI Experimentation: Act as a 'zero-to-one' builder. You stay current with the latest trends in LLMs, RAG (Retrieval-Augmented Generation), and Agentic frameworks (like LangGraph or AutoGen) to rapidly prototype and validate new 'System of Action' capabilities
• Hands-on, High-Leverage Contributor: You ship production code. You will lead the development of high-impact initiatives, moving from a Jupyter notebook prototype to a petabyte-scale production system with ease
• System Reliability & Trust: Ensure that GenAI outputs are not just 'cool,' but trustworthy. You will drive the strategy around LLM evaluation, observability, and guardrails to ensure accuracy and data privacy
• Technical Mentorship & Multiplier: You will elevate the AI fluency of the entire engineering org. You'll mentor senior engineers on prompt engineering, model selection, and the nuances of building non-deterministic systems
• Cross-Functional Influence: Partner with Product and AI Research teams to scope opportunities where GenAI can solve real-world GTM friction, ensuring our technical bets align with HubSpot's broader AI strategy
• Rapid Domain Fluency: The AI landscape changes weekly. You must be able to dive into a new paper, model, or framework today and explain its strategic value to the team tomorrow
• User-First Mindset: You have a track record of building products that people love. You advocate for the end-user in design reviews and are willing to pivot technical strategies if it means a better outcome for the customer
• Proven Expertise in Distributed Systems: You have experience building and scaling high-throughput, low-latency systems with strict reliability requirements
• GenAI Proficiency: Deep experience with LLM orchestration, vector databases, and building production-grade RAG pipelines. You understand the trade-offs between different model architectures (GPT-4o, Claude, open-source models)
• Agentic Thinking: A background in building autonomous or semi-autonomous systems where AI takes actions, handles tool-calling, and manages complex states
• A Bias for Action: You prefer a working prototype over a 20-page design doc. You have a habit of turning ambiguity into clear plans and working software
• Product Intuition: You care about the 'why' behind the 'what.' You understand the pain points of sales and marketing professionals and build technology that solves them