Intravenous infiltration is a condition wherein an infused solution leaks inadvertently into soft tissue surrounding a hypodermic needle site. This occurrence affects approximately 6.5% of patients in hospitals worldwide, and can lead to severe tissue damage if not treated immediately. The methods currently used by medical staff to detect an infiltration are subjective and can potentially be prone to error. Infiltration is an even larger concern in pediatric patients, who have smaller veins than adults and have more difficulty in communicating pain or other discomfort associated with the infiltration with medical staff. For these reasons, automatic IV infiltration detection could potentially reduce the risk associated with this damaging condition. This paper proposes a novel proof-of-concept system that uses non-invasive sensing in conjunction with a low-power embedded computing platform to deliver continuous infiltration monitoring around the IV catheter site. This kind of system could be able to detect an infiltration by non-invasively monitoring for known symptoms: swelling of soft tissue and increased skin firmness; these symptoms can be sensed by measuring skin stretch and local bioimpedance. Moreover, the low-power design and wireless capabilities can potentially enable continuous wear. The proposed automatic IV infiltration detection system could significantly improve the number of infiltrations identified and treated on time.