MISRAM LLC

AI/ML is explicitly named and the core deliverable (trusted AI for OPIR automation) suggests genuine ML work, likely anomaly detection or sensor fusion, but 'trusted AI' framing suggests significant effort on interpretability/validation infrastructure alongside model development, lowering it from 8-9. Problem is operationally grounded (overhead persistent infrared battlespace awareness) but lacks specificity on failure modes, user workflow, or success metrics. MISRAM LLC is an unfamiliar name without clear legacy-integrator signals, but the LLC suffix and lack of 'Systems/Solutions' language prevents a higher score absent evidence of known AI-native pedigree.