Led full-stack engineering for TKCORE AI, a context-aware AI workspace.
Key Responsibilities & Achievements:
• Signature Model Integration: Handled optimization and core execution pipelines for our flagship model, TkCore-V5.5-Pro, ensuring high-performance reasoning and robust multimodal capabilities.
• Context-Aware RAG Pipelines: Engineered vector data ingestion systems enabling users to upload custom knowledge base files, grounding the AI instantly in localized project backgrounds.
• Multi-Model Aggregator: Built a scalable orchestration layer to seamlessly unify access to top-tier global models like DeepSeek-V4, Qwen-Max, and GLM-5.1 within a single dashboard.
• End-to-End Workspaces: Architected an isolated project management environment linking data-grounded insights directly to creative generation tools (AI music, writing, and visual assets).
Gained deep expertise in scalable LLM middleware design and state-management for complex retrieval systems.
