DSML.KZ Community Hub

Eventum AI is an AI engineering consultancy building production-grade LLM, RAG, agent, and document-intelligence systems. Responsibilities: • Lead architecture and hands-on implementation of a production RAG / knowledge platform • Build document ingestion for Google Workspace, PDFs, scans, tables, and images • Implement permission-aware retrieval with ACL enforcement at query time • Build hybrid search, embeddings, reranking, and citation-backed generation • Deploy and benchmark open-weight models using vLLM or similar infrastructure • Build evaluation/regression systems for retrieval, answer quality, security, and model upgrades • Own production deployment, observability, CI/CD, documentation, and handoff • Work directly with Eventum’s senior team and client technical leadership Requirements: • 6+ years of professional software/ML engineering experience • Strong experience building production LLM/RAG systems • Excellent Python/backend engineering skills • Hands-on experience with embeddings, vector search, reranking, and document pipelines • Experience serving open-weight LLMs with vLLM, SGLang, Triton, TGI, or similar • Strong cloud/infrastructure skills; Azure, Docker, Terraform/Kubernetes, and CI/CD are especially relevant • Experience with LLM evaluation and production reliability • Comfortable owning ML, backend, data, and infrastructure rather than working in a narrow specialty Optional: • Google Workspace APIs and permissions • pgvector/Qdrant • BM25 + dense retrieval • OCR/document AI • multimodal models • Prometheus/Grafana • no-egress or security-sensitive environments