PinnedInProsus AI Tech BlogbyAsad Ismail·Mar 9MemEval: Benchmarking Memory for AI AgentsAs AI agents move from single session chatbots to persistent, multi-session assistants, memory has become a critical capability. Comparing…A response icon1A response icon1
InProsus AI Tech BlogbyAsad Ismail·Sep 15How we evaluate ToqanWhat we learned from six months of evaluating ToqanA response icon3A response icon3
InProsus AI Tech BlogbyAsad Ismail·Aug 11Scaling Open-Weight Inference for Long-Context AgentsOpen-weight models are ready, but scaling them in production is the hard part.
InProsus AI Tech BlogbyAsad Ismail·Jun 4Beyond Prompt Optimization: Optimizing the Whole Agent HarnessMost agent optimizers tune the prompt. Yet in modern agentic systems, orchestration, tool usage, runtime behavior, and execution policies…