Agentic AI Engineer
Team: Federated Data Lake
Project: BRAIN++ – Bulgaria’s AI Factory Location: Sofia Tech Park (Hybrid / On-site) Type: Full-time
1. About the Project and Our Mission
As part of BRAIN++, Bulgaria’s first comprehensive AI Factory, we are building the core infrastructure that will support the development and deployment of national foundation models. The initiative aims to strengthen Bulgaria’s AI ecosystem and contribute to European digital sovereignty through the development of advanced AI capabilities, governance frameworks, and cross-sector AI applications.
Within this initiative, the Federated Data Lake team is building a highly secure federated data infrastructure powered by an autonomous agent ecosystem. The platform enables trusted collaboration between institutions, researchers, public sector organizations, and industry partners while ensuring strict compliance with European data governance and the EU AI Act.
To kickstart this ecosystem, we are developing the first autonomous agents responsible for critical workflows, including:
● a multimodal data anonymization agent capable of securely processing text, video, and audio datasets
● a legal-contract-to-ODRL translation agent that converts legal agreements into machine-readable policy frameworks for automated data governance
We are looking for engineers who thrive on solving unprecedented architectural challenges across multiple technical domains. We do not measure candidates by years of experience, but by their ability to engineer reliable, production-grade AI agents from cutting-edge foundational models.
2. Role Overview
We are seeking an Agentic AI Engineer to design, build, and deploy the first generation of
autonomous AI agents operating within the Federated Data Lake ecosystem.
The role focuses on transforming foundation models across text, vision, and audio into robust autonomous workers capable of executing complex workflows involving sensitive data processing, governance enforcement, and distributed system integration.
You will work at the intersection of AI engineering, distributed systems, and secure data infrastructure, building agents capable of operating reliably in high-stakes environments such as government services, healthcare, and research collaborations.
The ideal candidate is a multidisciplinary engineer capable of architecting entire systems from scratch — from data pipelines and model orchestration to deterministic output validation and distributed training integration.
3. Required Qualifications
Demonstrated Architectural Mastery
Proven ability to architect, deploy, and monitor production-grade LLM applications and autonomous agentic workflows from scratch. You must understand state management, tool orchestration, and agentic design patterns at a fundamental level.
Elite Problem-Solving Capabilities
Exceptional ability to debug and synthesize solutions across highly ambiguous and cutting-edge technical stacks. You must be comfortable entering unfamiliar domains and mapping scalable architectural solutions rapidly.
Multidisciplinary Systems Engineering
You are fundamentally a backend systems engineer, capable of integrating large-scale machine learning models into secure, high-throughput distributed infrastructures.
Strict Deterministic Output Enforcement
Proven ability to enforce deterministic outputs from probabilistic models, ensuring compliance with strict schema definitions and formal semantic ontologies such as JSON-LD, XML, or ODRL.
Zero-Tolerance Data Pipelines
Hands-on expertise building robust multimodal data anonymization pipelines, including:
● Computer vision workflows for video redaction and blurring
● Audio processing pipelines for voice masking
● Advanced NER-based text anonymization for sensitive information
A strict zero-tolerance approach to data leakage is essential.
AI Training Pipeline Integration
Experience integrating distributed storage systems with AI training pipelines, enabling large-scale datasets to stream directly into PyTorch, TensorFlow, or custom training jobs without starving compute resources.
4. Preferred Qualifications
● Familiarity with Open Digital Rights Language (ODRL) or other semantic web rights-expression frameworks.
● Experience in LegalTech, RegTech, healthcare, or financial data environments
where handling highly sensitive information is required.
● Experience deploying containerized ML pipelines using technologies such as Docker and Kubernetes, particularly for processing large-scale video and audio datasets.
● Experience working with federated data infrastructures, data governance frameworks, or distributed AI systems.
5. What We Offer
● Opportunity to work on BRAIN++, Bulgaria’s flagship AI infrastructure initiative.
● Participation in building the core Federated Data Lake powering national AI foundation models.
● Work on cutting-edge agentic AI systems and autonomous data processing architectures.
● Flexible working environment
● Opportunities for professional development and research collaboration
● Access to advanced computing infrastructure and large-scale AI datasets
Deadline: April 8, 2026
More information and application: [email protected]
Project No. 101250707 – BRAIN++ is implemented under a funding agreement with the European High-Performance Computing Joint Undertaking and the European Commission.
Sofia Tech Park is Bulgaria’s first science and technology park – a hub for researchers, start-ups, and established technology companies, supporting the development of innovation in the fields of ICT, life sciences, and clean energy.