Evinced – GenAI Engineer

Evinced is on a mission to make digital accessibility universal. Join as a GenAI Engineer in Herzliya to build and enhance AI infrastructure that detects and resolves accessibility issues through multimodal LLM pipelines and developer-facing tools.

Responsibilities:
– Design and maintain production-grade RAG pipelines and LLM-based systems
– Fine-tune models using LoRA or QLoRA to boost accuracy
– Innovate evaluation strategies to maintain high-quality GenAI outputs
– Collaborate with product, backend, and DevOps teams to integrate AI into SaaS accessibility tooling
– Troubleshoot across services and proactively resolve pipeline issues

Requirements:
– 3+ years with RAG pipelines and LLM APIs
– 5+ years working with NLP/LLM systems in production
– Experience fine-tuning (LoRA/QLoRA) and using Python libraries (LangChain, numpy, scipy)
– Proficient with cloud platforms (GCP, AWS, or Azure)
– Advantageous: multimodal LLM experience, MCP servers, and working with scalable data stores (ElasticSearch, BigQuery)