AAI Labs
AAI Labs is a Lithuanian AI engineering and research company that builds custom, applied AI systems for businesses, public institutions and governments. Its work spans LLMs, speech recognition, computer vision, credit-risk modeling, data analytics and process automation, with an emphasis on reliable, interpretable and EU-compliant systems.
At a Glance
- Businesses and enterprises
- Public institutions and national/local governments
- Manufacturing and engineering
- Healthcare and medicine
- +6 more
AI Tools by AAI Labs
(1)
Agent Barn
Open Source Kubernetes Agent Control Plane
Discussions
No discussions yet
Be the first to start a discussion about AAI Labs
Latest News
AAI Labs published “How not to be an AI slop programmer.”
AAI Labs published “AI Agents in Business: How Working with Software Will Change,” describing bounded, verifiable agent workflows and human oversight.
AAI Labs published “How AI Scan Finds Where AI Fits Your Business,” explaining its evidence-based process-discovery and prioritization product.
AAI Labs published its technical field note on an AI code-reviewer agent operating through Slack, Bitbucket and GitHub while leaving merge decisions to humans.
Products & Services
Dedicated, pre-vetted AI/ML engineering teams assembled for a client challenge, with Starter, Build, Scale and Configure packages, monthly scaling, on-premise or VPC deployment options, and client ownership of foreground IP.
A structured assessment that interviews an organization's people and produces an evidence-based, ranked report of where AI could create value; it is intended to identify priorities rather than build the implementation itself.
An open-source orchestration layer/control plane for deploying, operating and governing fleets of AI agents on a company's own Kubernetes infrastructure. It supports runtimes such as OpenClaw and Hermes, reusable templates and skills, integrations, lifecycle management and observable activity.
Chatbots and conversational AI for regulated industries.
Market Position
AAI Labs positions itself between generic AI software vendors and large global consultancies: it offers smaller, rapidly deployable engineering teams and custom systems for messy, domain-specific workflows, with European GDPR/EU AI Act orientation and on-premise/VPC options. Agent Barn differentiates through self-hosting, open source, Kubernetes-native operation, inspectable code and supervised governance; relevant alternatives include commercial agent platforms such as Microsoft Copilot Studio, Salesforce Agentforce and ServiceNow, and open/self-hosted projects such as LangGraph/LangSmith, Dify and CrewAI.
Leadership
Founders
Aistis Raudys
Founder and CEO; Dr. Raudys holds a PhD from Vilnius University in multivariate data analysis, teaches AI at Vilnius University, and has worked as a quantitative researcher and in algorithmic trading with Deutsche Bank, BNP Paribas and Société Générale.
Tadas Šubonis
Co-founder and technical founder; an AI/software architect with an MSc in Artificial Intelligence and a BSc in Physics. His experience includes NLP news readers, neural speech systems, mobile insurance and credit-scoring platforms, voice AI, full-stack engineering and system architecture.
Executive Team
Aistis Raudys
Founder & CEO
PhD in AI/multivariate data analysis from Vilnius University; AI professor and researcher with experience in algorithmic trading, automated funds and quantitative research at Deutsche Bank, BNP Paribas and Société Générale.
Tadas Šubonis
Co-founder / Technical Founder
MSc in Artificial Intelligence and BSc in Physics; specializes in system architecture, voice AI, full-stack engineering and applied AI, with more than a decade of ML and software-architecture experience.
Founding Story
AAI Labs grew out of Aistis Raudys and Tadas Šubonis reconnecting through CoderDojo, where Raudys taught children to code. Their first major collaboration was a voice-driven virtual assistant for real-estate company Hanner; the project led to further partnerships, including work for Factris, and developed into a full company. Their initial vision was to make applied AI useful in real operating environments rather than build technology demonstrations in isolation.
Business Model
Revenue Model
AAI Labs sells custom AI engineering, research-and-development and consulting services, dedicated engineering teams on monthly packages, and productized AI solutions. Agent Barn adds paid supervised pilots and managed cloud operation alongside a free self-hosted open-source edition; customers pay for model inference separately in the cloud offering.
Pricing Tiers
Concept validation and small-scale projects; 1.5 FTE.
Production development; 3 FTE.
Enterprise-level operations; 4 FTE.
Specialized team for unique or complex solutions; tailored effort.
Planned Telegram personal agent with email triage and competitor research; includes $20 of monthly usage.
Managed hosting, upgrades, organizations, roles and agent lifecycle controls; customer still pays model-inference costs.
One workflow scoped end-to-end with hands-on support, controls review and handover.
Target Markets
- Businesses and enterprises
- Public institutions and national/local governments
- Manufacturing and engineering
- Healthcare and medicine
- Transport and logistics
- Energy
- Manufacturing quality control, production coordination and waste reduction
- Healthcare documentation and medical-language-model applications
- Transport demand forecasting, route planning and fleet optimization
- Public-sector and municipal process automation
- Energy-sector analytics and optimization
- Call-center automation, voice assistants and speech analytics
- Hanner
- Factris
- JUDU / Vilnius public transport
- Kautra