14 ott - Italia
Point-out
pPoint-out is hiring an bAI Engineer /b to build the agentic AI workflows at the heart of our platform. You'll take agents from prototype to production and keep them measurably good, working side by side with our founder CTO. /p h3About Point-out /h3 pEvery day, millions of people ask AI what to buy, whom to believe and how to stay informed. AI assistants' answers have become a new medium where knowledge takes shape and choices are guided. We want to understand how that happens. /p pWe're a team born of the friction between physics, economics and communication. We build intelligence products that measure, and make legible, how the media, AI models and AI agents select their sources and portray companies and people. /p pOur platform for GEO (generative engine optimization) and GEA (generative engine advertising) measures how a brand and its competitors appear in AI searches on ChatGPT, Gemini, Perplexity and Claude. It identifies what to do about it across owned channels, earned media and paid investments, and helps marketing and communication teams act within the tools they already use. We're building from Italy with expansion to other European markets planned. We recently closed our first funding round and are preparing the next. /p h3The role /h3 pYou'll design, build and operate the agentic workflows behind our most advanced use cases: /p ul lidigital twin simulation of AI searches; /li licomplex classification; /li licrawlability audits; /li lieditorial recommendations; /li liMCP integrations with general-purpose AI assistants. /li /ul pThis is a hands-on, junior to mid-level role. The CTO will pair with you on the hardest problems, and you'll take ideas from prototype to a running,
monitored service with growing autonomy. /p h3What you'll do /h3 ul liTurn platform use cases into precise requirements: inputs, outputs, quality criteria, cost and latency constraints. /li liScout and test models, frameworks and managed services on our real tasks and data, and recommend what to adopt, watch or ignore. /li liDesign and prototype multi-step agents that combine LLM reasoning, tool use, retrieval and structured outputs. /li liBuild evaluation sets and metrics; measure output quality, token cost, latency and failure modes; decide with the CTO when a solution is ready to ship. /li liDeploy agents as services on AWS (Amazon Bedrock AgentCore and containers) and integrate them through clean, documented APIs. /li liInstrument every agent with tracing and evaluation (Langfuse) and catch data and quality drift before it reaches customers. /li liDocument architecture decisions, prompts, evaluation results and runbooks so the next engineer can pick up where you left off. /li liAs labelled data grows, train supervised classifiers (scikit-learn, PyTorch) to replace generative models where they're faster, cheaper and more predictable. /li /ul h3What we're looking for /h3 ul libSolid Python engineering /b: clean, tested, version-controlled code (Git/GitHub), sound architecture and API design.
You're confident with pandas and familiar with scikit-learn and PyTorch. /li libData and SQL /b: you can model, query, join and shape data, and you check data quality before model quality /li libAgent frameworks /b: you've built agentic or RAG workflows with LangChain/LangGraph, LlamaIndex, Strands Agents or similar, and you understand prompting, tool use, memory, guardrails, structured outputs and typical LLM failure modes /li libObservability and evaluation /b: you can design evals, define metrics and use tracing tools like Langfuse to understand what an agent actually did and why /li libCloud and deployment /b: you know Docker and have deployed services on AWS or another cloud, getting from your laptop to a running, monitored service /li libAI-native way of working /b: you build with tools like Claude Code and judge their output critically /li libClear communication in Italian and English /b, with technical and non-technical colleagues alike /li /ul h3Nice to have /h3 ul liAmazon Bedrock AgentCore, Bedrock model APIs or other managed agent runtimes; AWS certifications. /li liInfrastructure as Code (Terraform, AWS CDK, CloudFormation), CI/CD pipelines, Kubernetes. /li liBuilding or integrating MCP servers with Claude, ChatGPT or Gemini. /li liSupervised text classifiers shipped to production. /li liWeb crawling, technical SEO or GEO. /li liExperience in an early-stage startup or small product team. /li /ul h3What we offer /h3 ul li€30 - 40k gross per year, depending on experience. /li li10% annual premio (50% on personal performance, 50% on company performance). /li liPermanent, full-time contract, fully remote. /li /ul #J-18808-Ljbffr
14 ott - Milano
Idro 2000
14 ott - Ciriè
Filippo
14 ott - Udine
Gruppo Danieli
14 ott - Nova milanese
New Age