A
Applied AI Engineer
·Israel
Not specifiedFull-timeAI EngineeringTechnologyInformation TechnologyInternet
Applied AI Engineer responsible for designing and developing multi-step AI agents, building RAG and retrieval systems, developing ground truth layers with source tracking and confidence scoring, and optimizing production LLM systems for accuracy, latency, and reliability.
Responsibilities
- Design and develop multi-step AI agents and workflows.
- Build RAG, retrieval, and context assembly mechanisms over large scale knowledge bases.
- Develop a Ground Truth layer incorporating source tracking, versioning, confidence scoring, citations, and contradiction detection.
- Build automated evaluation systems, regression tests, and quality gates for AI generated outputs.
- Process and extract intelligence from PDFs, Word documents, Excel sheets, presentations, images, and complex templates.
- Optimize for accuracy, latency, reliability, and inference costs.
- Investigate production failures and translate them into systemic tests and improvements.
Requirements
- Deep understanding of KV cache management, top-k/top-p sampling, and context window optimization.
- Proven ability to deploy, serve, and run open weight AI models locally, including hardware requirements like GPU vRAM calculations and quantization strategies.
- Hands-on experience with model fine-tuning, specifically PEFT methods like QLoRA, and optimization frameworks such as Unsloth.
- Solid theoretical foundation in neural network architectures, including Transformers, RNNs, GRUs, and CNNs.
- Proven experience building LLM-based systems beyond simple prompt engineering (model APIs, structured outputs, function/tool calling, agentic workflows).
- Deep understanding of RAG, embeddings, retrieval, and vector search.
- High proficiency in Python.
- Ability to build production grade systems including testing, logging, monitoring, and fallback handling.
Nice to have
- Experience with AWS, Azure, GCP, or OCI, with familiarity with Terraform or Infrastructure as Code (IaC).
- Extra experience in document intelligence, OCR, multimodal models, knowledge graphs, hybrid search, or reranking techniques.
- Familiarity with MCP (Model Context Protocol) and multi-agent systems.
- Experience building Enterprise B2B products with a focus on security and tenant data isolation.