Senior Solutions Architect

Directeam helps startups scale on AWS efficiently, and increasingly helps them build with AI. As an AWS Premier Partner and an Anthropic partner, we combine our flagship automated savings engine, Sombrero Savings, with our latest breakthrough: a Real-Time AI FinOps Agent. We don't just manage costs, we provide a strategic advantage that converts cloud waste into business runway.

Our AI-backed service has already been adopted across our client base, delivering an average monthly cost reduction of more than 30%. We are bootstrapped, profitable and growing fast, and we work with leading Israeli tech companies across cyber, fintech, sports-tech, AI and SaaS, from pre-seed to growth stage.

We're looking for a Senior Solutions Architect (AI-Native Cloud) to be the trusted technical advisor for a portfolio of startups: designing their cloud and AI architectures, helping them run those architectures reliably and cost-efficiently, and putting generative AI and agents into production, not just into demos. About 80% of the role is customer-facing. This is a hands-on role working alongside a small team of senior architects, our Account Managers, and AWS and Anthropic field teams.

What You'll Do

• Own the technical relationship with your customers. Lead discovery and architecture review sessions, turn business goals into concrete AWS designs, and act as the escalation point for their hardest production problems.
• Design and build cloud-native architectures, from greenfield to modernization, that are highly available, secure and cost-effective, using the AWS Well-Architected Framework, including the Generative AI Lens.
• Architect GenAI and agentic solutions on AWS: RAG and knowledge systems, AI agents and multi-agent workflows, and LLM-powered product features, using building blocks such as Amazon Bedrock (including Claude models), Bedrock AgentCore, Knowledge Bases and Guardrails, vector stores (OpenSearch, pgvector, S3 Vectors) and MCP-based tool integrations.
• Take AI from POC to production. Define evaluation, observability, security and governance for LLM workloads (prompt-injection risk, data privacy, IAM and least privilege for agents), and keep inference cost and latency under control.
• Bring the FinOps mindset to everything, including AI. Run cost-optimization sessions, right-size compute and inference (model selection, prompt caching, batch, provisioned throughput), and connect customers to Sombrero Savings and our AI FinOps Agent.
• Work AI-native yourself. Use AI coding and agent tools such as Claude Code, Kiro or Amazon Q Developer to speed up IaC, reviews, migrations, troubleshooting and documentation, and help customers' engineering teams adopt the same practices.
• Write SOWs and plans that hold up. Scope work realistically, separate what matters from what doesn't (80/20), and hand over cleanly to delivery teams.
• Drive growth with the account team. Spot opportunities proactively, including AI use cases, modernization, migration and AWS funding programs, before the customer has to ask.
• Build thought leadership. Write blogs, run workshops, webinars, meetups and hands-on AI enablement sessions, and turn repeated field solutions into reusable playbooks, Terraform modules and reference architectures.
• Feed the product. Bring field insight back into our FinOps products and internal AI tooling, and influence how they evolve.

• 5+ years with cloud technologies, including 3+ years designing and running production systems on AWS.
• 3+ years as a tech lead or architect for large-scale, production-grade products, ideally in a customer-facing role (Solutions Architect, Cloud Architect, consulting).
• Proven expertise in designing, analyzing and troubleshooting large-scale distributed systems, including networking, security best practices and Linux.
• Hands-on GenAI experience on AWS: you have built at least one LLM-based solution that reached real users (RAG, agent or AI feature), preferably with Amazon Bedrock, and you understand its architecture, cost and failure modes.
• Working knowledge of agentic patterns: tool use and function calling, MCP, orchestration, memory, guardrails and evaluation.
• Software development skills in Python and/or TypeScript, Go or Java. Read and write code, not only diagrams.
• Infrastructure as Code (Terraform or CDK) and CI/CD.
• Daily use of AI-assisted development tools (e.g. Claude Code, Kiro, Cursor, Amazon Q Developer), with good judgment about when to trust them.
• Excellent verbal and written communication in Hebrew and English.

Advantages

• Kubernetes / EKS and its ecosystem (Karpenter, KEDA, ArgoCD), and serverless architectures.
• Data platforms and databases (SQL, NoSQL, data lakes; Glue, Athena, Redshift, OpenSearch).
• FinOps experience: CUR analysis, Savings Plans / RIs, cost-aware architecture, AI inference cost modeling.
• Amazon SageMaker, model fine-tuning, or self-hosted open-source models on GPU (vLLM, Inferentia/Trainium).
• AI observability and evaluation tools (e.g. Langfuse, LangSmith, Bedrock evaluations) and agent frameworks (Strands Agents, LangGraph, Claude Agent SDK).
• Certifications: AWS Solutions Architect Professional, AWS Machine Learning / AI Practitioner, Claude Certified Architect.
• Experience designing SaaS / multi-tenant platforms, and a good understanding of the startup ecosystem.

Our office is located at Beit Geibor Sport Tower, Ramat Gan, near Tel Aviv Savidor Central railway station.