Aakriti Aggarwal
AI research engineer building reliable systems and runtime cost control for LLMs.
Aakriti Aggarwal builds reliable AI systems and optimizes model inference at IBM Research. Also a Microsoft MVP, she focuses on making enterprise models practical rather than just experimental, sharing her findings through IEEE papers and technical blogs.
Currently, she develops tools to catch infinite loops and token bloat before they ruin engineering budgets. Her recent work dives into building runtime cost control planes for LLMs, showing teams how to enforce policies at the SDK level and prevent runaway agent spend.