AI Engineer | Multimodal AI | LLM | RAG | Retrieval Augmented Generation | LangGraph | CrewAI | Agents SDK

AI Engineer / Researcher

Your Mission

You'll be a key contributor translating cutting-edge multimodal AI research into robust, production-grade agentic systems. By integrating vision models, large language models (LLMs), and advanced retrieval pipelines, your work will power real-world automation workflows and significantly boost productivity across complex domains.

You'll be working at a well-funded, early-stage venture led by a technical founding team with deep expertise in AI, engineering, and startup execution. They value engineering rigor, scientific innovation, and building impactful products that solve real operational problems.

What You’ll Do

1. Agentic Systems
   • Design, build, and refine agentic systems using frameworks like LangGraph.
   • Integrate LLM-based reasoning with multimodal models and domain-specific tool use.
   • Rapidly transition validated concepts into reliable, production-ready solutions.


2. Data Retrieval and Organization
   • Build and optimize multimodal data retrieval pipelines.
   • Work with structured and unstructured documents such as technical drawings, plans, and visual datasets.


3. Evaluation
   • Offline Evaluation: Develop high-quality benchmarks and use-case-specific evaluation metrics.
   • Online Evaluation: Monitor live system performance and implement learnings for continual improvement.

Must-Have Qualifications

• 5+ years of experience building production-grade ML systems OR 2+ years in a top-tier AI startup focused on agentic or multimodal systems.
• Expert Python development skills.
• Experience with agentic frameworks like LangGraph, CrewAI, or Agents SDK.
• Demonstrated ability to translate complex AI research into scalable, real-world systems.

Nice-to-Have Qualifications

• Familiarity with technical or visual domain-specific data types (e.g., CAD, IFC, point clouds).
• Experience with vector search technologies (FAISS, Weaviate, Pinecone, etc.).
• Previous work in an early-stage startup with fast iteration cycles.
• Experience training or fine-tuning cutting-edge multimodal or vision-language models.