Razi Ashary

Head of Research and Development @Airs Foundry

SG

AutoDev Development: Architected and currently developing AutoDev, a model-agnostic downloadable AI developer agent platform purpose-built for the AIRS Foundry worker pipeline. Engineered to support full agentic capability across any LLM provider via a single binary, including Skills, Recipes, Subagents, MCP servers, and Hooks, with zero vendor lock-in. Core Logic Architecture (AIRS Forge): Designing and maintaining AIRS Forge, the central intelligence layer of AIRS, comprising the core-orchestrator backend, memory layer, and supporting microservices. Active development includes an end-to-end speech-to-speech translation system, a Node-Based Builder for API chaining, and a neuroscience-grounded memory architecture with a chemistry-inspired structural design. Blueprinting & System Innovation: Designed novel system architectures across multiple product verticals, including a proprietary agent-to-agent (A2A) protocol, an LLM universality layer, and a periodic-table-inspired AI memory system, establishing the technical direction for AIRS product development. Cross-Departmental Technical Guidance: Provided ongoing technical direction to 3 departments including manufacturing, marketing, and operations in areas of AI integration, system architecture, and innovation strategy, acting as the primary technical reference across the company. Cloud Infrastructure & Deployment: Managed GCP infrastructure including Cloud Run services, GCE VM configurations, S3 and GCS bucket migrations, and cross-account deployments for AIRS product environments, ensuring stable and scalable multi-service operations. Research & Knowledge Management: Conduct weekly research on AI models, tools, frameworks, and papers, producing monthly findings documents with applicability notes for company integration and advancing selected findings into active prototypes or formal proposals.

Institut Teknologi Bandung

Bachelor's degree