Blog
Thoughts, stories and ideas about software development, design, and more.
Giving Embodied Agents Memory: How Graph Databases Enable Smarter Long-Horizon Planning
Exploring how integrating graph-based episodic memory into embodied agents improves long-horizon task completion, reduces action repetition, and enables learning from past mistakes in complex household environments.
Teaching Robots to See and Act: Building a Vision-Language Model for Precise Manipulation
Exploring how combining 3D perception, depth estimation, and visual in-context learning enables embodied agents to achieve 71.67% accuracy on complex manipulation tasks, outperforming state-of-the-art multimodal language models.
Teaching LLMs to Reason About Finance: A Deep Dive into FinQA
Exploring how large language models can understand financial reports through symbolic reasoning, comparing in-context learning with LoRA fine-tuning, and integrating external tools for reliable quantitative analysis.
Building the Future of Autonomous Trading: From User Intent to Agent Execution
Exploring how the Agentic Trading Platform transforms cryptocurrency trading through intelligent agents that operate autonomously with user-defined strategies and risk parameters.
Agent Commerce Protocol: Operari joins Autonomous Commerce
How Operari's Computer-Using Agents are now participating in the Agent Commerce Protocol network, enabling autonomous agents to request DeFi swaps through natural language interactions.
Operari: Revolutionizing DeFi with Computer-Using Agents
Exploring how Computer-Using Agents (CUAs) are revolutionizing DeFi interactions by combining conversational AI with visual interface capabilities.