• Design and implement multi-agent architectures using LangGraph • Build stateful, tool-augmented agents capable of reasoning, planning, and execution • Develop agent orchestration patterns (planner-executor, supervisor, hierarchical agents) • Implement memory systems (short-term, long-term, contextual memory) • Enable cross-agent communication and coordination • Build production-grade LLM applications (RAG, agents, copilots) • Integrate tools, APIs, and enterprise systems into agent workflows • Design robust prompting, routing, and guardrails • Optimize workflows for latency, cost, and quality • Develop reusable AI services, agent frameworks, and APIs • Deploy agent-based systems on AWS • Use Docker and Kubernetes for scalable deployments • Implement monitoring and observability (LangSmith) • Evaluate agent performance across reasoning quality and task completion • Improve workflows using feedback loops and tracing • Diagnose and fix failure modes in multi-step pipelines • Work with product and engineering teams to build agentic workflows • Guide teams on best practices for agent design • Contribute to reusable patterns and frameworks

• Design and implement GenAI solutions using AWS Bedrock and Agentcore. • Define architecture for LLM-based applications, including RAG pipelines and agentic workflows. • Develop and orchestrate agentic AI workflows, enabling multi-step reasoning, tool usage, and task automation. • Build and manage RAG pipelines, including embeddings, retrieval mechanisms, and vector databases. • Integrate LLM capabilities into enterprise applications via APIs and backend services. • Design and optimize prompt engineering strategies for accuracy, relevance, and performance. • Work with structured and unstructured data sources to enable knowledge-driven AI applications. • Ensure model evaluation, monitoring, and optimization for latency, cost, and response quality. • Collaborate with application, data, and platform teams for end-to-end solution delivery. • Define best practices for security, governance, and responsible AI usage. • Troubleshoot and resolve issues in production GenAI systems. • Provide technical leadership and mentor team members while remaining hands-on.

What we do Snezzi helps brands show up in AI answers on ChatGPT, Perplexity, Google AI Mode, and others. We handle everything end to end: audits, content, technical fixes, backlinks, and measurement. Fully done for you. We're a small team of second-time founders. We ship weekly. We run AI agents for most of our internal work. And we're growing fast enough that we need sharp people who want to learn this space by doing real work, not by watching from the sidelines. What this internship actually is You'll work on live client accounts from week one. That means real audits, real fixes, real reporting. Not shadowing. Not slide decks. Not busywork. By the end of the engagement, you'll have shipped actual work, built skills that are genuinely hard to find right now, and have a portfolio that proves it. Strong performers convert to full-time. We'll tell you early if we see that path for you. What you'll do - Run AI visibility audits across ChatGPT, Perplexity, and Google AI Mode - Help identify gaps and assist with fixes: content updates, schema, technical cleanup - Use Claude Code and other AI tools to build and run scripts (no prior coding required — you learn by doing) - Help prepare client reports, weekly updates, and pre-reads - Document and templatize every process you touch so the next person can use it Tools you'll use Claude Code, ChatGPT, Perplexity, Google Search Console, GA4, Ahrefs, Shopify, WordPress. You'll learn them on the job. What matters is how fast you pick things up. Who we're looking for You don't need a specific degree or years of experience. You need the right instincts. Must-haves: - You use AI tools a lot, out of curiosity, not just to finish assignments - You write clearly. Most of this job is communication: with clients, with AI agents, with your team. - A terminal or a code file doesn't scare you. You don't need to write code, but if Claude generates a script, you can run it, read the output, and ask Claude to fix what broke. - You try before you ask. "I don't know" is a starting point, not an excuse. - You want to ship things, not just observe them. Moves you to the top of the pile: - You've built or shipped anything on the internet: a blog, a store, a tool, a side project, a newsletter - You've poked around GA4 or Search Console before - You've written content for a brand or website - You've done any kind of SEO, growth, or content work Not a fit if: - You think prompting an AI handles everything. Taste and judgment still matter. - You want clean handoffs and tightly scoped tasks - Client-facing work makes you uncomfortable

What we do Snezzi helps brands show up in AI answers - on ChatGPT, Perplexity, Google AI Mode, and others. We handle everything: audits, content, technical fixes, backlinks, and measurement. Fully done for you. We're a small team of second-time founders. We ship daily. We use AI agents for most of our internal work. And we're growing fast enough that we need someone to help us deliver. The role You'll work directly with clients to improve how they show up in AI answers. You'll run audits, fix what's broken, publish content, and report what moved. This is not a ticket-taking job. You own accounts. You spot problems. You ship fixes. The week you join, you're already working on a real client. Most of your leverage comes from knowing how to use AI tools well — not from writing code from scratch. If you can take a Claude-generated script, run it, read the output, and ask Claude to fix what broke, you'll thrive here. What you'll actually do day to day: - Audit brand visibility across ChatGPT, Perplexity, and Google AI Mode - Identify gaps, map them to fixes, and ship them — content, schema, technical cleanup - Run client updates, prep pre-reads, and translate client questions into action - Turn every repeated task into a script or workflow so you never do it manually twice - Use tools like DataForSEO, Google Search Console, Ahrefs, and GA4 to build reports that actually mean something Tools you'll use: Claude Code, ChatGPT, Cursor, Perplexity, v0, Google AI Mode, Python/Node scripts (run and edit, not write from scratch), Shopify, WordPress, Webflow, GA4, Search Console, Ahrefs. You don't need to know all of these on day one. You need to pick things up fast. Who we're looking for Must-haves: - You use AI tools obsessively. You have strong opinions about ChatGPT vs Claude vs Perplexity. - You write clearly and edit ruthlessly. Communication is half this job. - You're not afraid of a terminal. If Claude Code generates a script, you can run it, read the output, and debug it with Claude's help. - You ship. Imperfect and out this week beats perfect and out next month. - You read docs. "I don't know" is a starting point, not a dead end. Moves you to the top of the pile (any one): - You've built something real with Claude Code, Cursor, v0, Lovable, or similar — even a toy project - You've shipped something on the internet: a blog, a store, a tool, a newsletter - You've done SEO, content, or growth work before - You've poked around GA4 or Search Console Not a fit if: - You think prompting an AI is a strategy on its own, without judgment or taste - You want clean scope, handoffs, and ticket queues - You avoid talking to clients