Staff AI Enablement Engineer

Vendelux

About The Position

Vendelux is transforming how companies discover, evaluate, and maximize the impact of events. Event marketers are the driving force behind pipeline and brand — yet events remain one of the least optimized and most opaque marketing channels. Vendelux changes that. We provide the system of record for event marketing, giving teams the data and insights they need to make smarter, more strategic decisions. Our AI-powered platform delivers proprietary insights across 250,000+ events, helping high-growth companies identify where their ideal customers will be, maximize ROI, and turn events into a scalable growth channel. Customers often describe Vendelux as an event marketer’s dream. A key part of this is our growing network of event organizer partnerships — where organizers share first-party attendee and sponsorship data, and in return gain access to valuable market insights. In addition, Vendelux Meetings helps customers turn event insights into action — using AI to identify high-value attendees and automatically book 1:1 meetings with the right prospects at conferences. This allows teams to maximize pipeline generation and make every event materially more impactful. Founded in 2021, Vendelux is a Series A SaaS company backed by leading investors including FirstMark, with a recent $14M round. Our team brings experience from companies like Bain, ZoomInfo, Shutterstock, Compass, Forter, Airbnb, and more. About the Role AI is no longer an engineering story at Vendelux. It's a company story. The way our sales team prospects, the way product researches, the way ops runs processes, the way data answers questions. All of it is changing fast. We're looking for a Staff AI Enablement Engineer to be the technical force behind that change. This role sits at the intersection of deep AI infrastructure and cross-functional impact. You'll build the MCP servers, agent pipelines, context frameworks, and orchestration layers that make AI genuinely useful across every team - not just for the engineers who've already figured it out on their own. The work is hands-on and technical. The impact is company-wide. If you're the person at your current company who got everyone on Claude Code, wired up MCP servers for internal tools, and is already running agents 24/7, you might be exactly who we're looking for.

Requirements

Strong software engineering fundamentals. You're building real infrastructure that teams depend on, not configuring existing tools
Deep hands-on experience with frontier AI agents (Claude Code, Codex, or equivalent) and the context engineering that makes them actually useful in complex codebases
Practical, production experience building with LLM APIs: tool use, multi-turn state, system prompt architecture, structured outputs, multi-agent orchestration
Hands-on experience with MCP or similar integration frameworks. You've connected agents to real production systems, not just toy examples
Experience designing for non-technical users: the agent that works for a software engineer is not the same as the one that works for a sales rep or an ops manager
Comfort working cross-functionally. You'll spend as much time talking to a head of sales or a product lead as you will writing code

Nice To Haves

Background in platform engineering, developer tooling, or data engineering
Experience with proactive/scheduled agent systems (not just request-response)
Familiarity with vector stores, RAG pipelines, or knowledge graph approaches for agent context
Experience with CI/CD automation involving AI agents
Exposure to B2B SaaS data models, CRM/MAP integrations, or event/attendee data

Responsibilities

Build the shared AI infrastructure layer
Design, build, and maintain MCP servers that connect our internal systems like Github, Snowflake, Linear, Notion, Slack, and others, to agents running across every function
Establish and own our context engineering standards: CLAUDE.md / AGENTS.md conventions, shared context/ directories, architecture docs that make our agents deeply aware of how Vendelux works
Build the memory and persistence layer for long-running agents: session continuity, proactive scheduling, cross-session context
Own orchestration infrastructure for multi-agent workflows: coordination, sub-agent spawning, token budgets, permission boundaries
Maintain codebase health as the system scales: shared component libraries, automated quality gates, fragmentation checks, doc validation in the PR pipeline
Drive adoption across every function
Partner with sales, ops, product, marketing, legal, and data teams to identify where AI can fundamentally change how a team works and build the agents that make it happen
Get every team to their “aha” moment fast: preconfigured environments, pre-connected tools, skills they can run immediately without debugging
Build and grow a skills marketplace where anyone can package a workflow and share it company-wide so one person’s breakthrough becomes everyone’s superpower
Create visibility and healthy competition around AI usage: leaderboards, showcases, Slack channels, all-hands demos that make building contagious
Identify force multipliers on every team (the people who get it early) and give them the platform and resources to bring their teams along
Build purpose-built agents for non-engineering teams
Each agent isn’t a chatbot. It’s a composition: the right MCP integrations, the right document access, the right memory system, the right workflows assembled into something that genuinely serves a function’s real work
Work closely with domain experts to turn institutional knowledge into something an agent can act on; the best agents are co-created, not handed down
Given Vendelux’s focus on event intelligence and pipeline, there’s particular leverage in agents that understand our data models, account scoring, and sales workflows
Own the hard infrastructure problems
Manage the access vs. safety tension: permissions scoping, token budgets, rate limiting, observability dashboards — guardrails that enable rather than block
Maintain reliability across agent infrastructure as the system grows: graceful degradation, fallback models, cost tracking
Evaluate frontier models, new MCP tooling, emerging agent frameworks, and integrate what's worth integrating before competitors catch up