
Technology
|
Public Safety Intelligence
|
YC W26
|
Valuation:
Undisclosed

Last Updated:
March 24, 2026

Builds a real-time AI-powered video intelligence platform that uses deep learning, multimodal vision-language models, and reinforcement learning to detect and escalate threats in live surveillance feeds for law enforcement and large-scale security operations.
Protent has publicly announced pilot deployments with police departments in Atlanta, Chicago, and St. Louis, focusing on real-time threat detection in live video feeds. Their YC W26 Demo Day profile highlights proactive surveillance intelligence, natural language video querying, and early escalation pattern detection as core platform capabilities.
The team's lean size (2-3 technical founders, no public job postings) signals a heads-down, pilot-driven development phase tightly coupled to law enforcement feedback loops. Founder Srihan Balaji's prior work deploying RL-optimized video intelligence at Lockheed Martin suggests proprietary reinforcement learning techniques for alert prioritization. Strong indicators of near-term expansion to adjacent verticals (stadiums, transit, corporate campuses) and likely development of integration APIs for existing VMS/PSIM platforms.
Real-time detection of weapons, fights, and medical emergencies in live surveillance video feeds using deep learning object detection and activity recognition models.
AI watches thousands of security cameras simultaneously and instantly alerts officers when it spots a weapon, a fight breaking out, or someone collapsing.
It's like giving every security camera its own tireless, eagle-eyed guard who never blinks, never takes a bathroom break, and immediately radios for backup the moment something looks wrong.
Natural language querying of live and archived surveillance video using multimodal vision-language models, enabling operators to search footage by describing events in plain English.
Instead of scrubbing through days of security footage, an officer can just type "person in red jacket near the east entrance at 2pm" and the system finds it instantly.
It's like having a Google search bar for your security cameras—except instead of searching the internet, you're searching everything every camera has ever seen, and it actually understands what you're asking.
Reinforcement learning-based alert prioritization system that dynamically ranks and filters threat detections to minimize false positives and surface the highest-risk incidents first for human review.
AI learns from every alert an officer acts on or dismisses, getting smarter over time about which threats are real and which are just someone waving a selfie stick.
It's like training a really smart assistant who learns that when your boss emails you at 11pm it's urgent, but when the office newsletter lands at 9am it can wait—except the stakes are a lot higher than inbox management.
Protent combines rare production experience deploying reinforcement learning for video intelligence in classified defense environments (Lockheed Martin) with cutting-edge UC Berkeley systems research, giving them a unique ability to build ML models that work reliably on messy, real-world surveillance feeds, not just academic benchmarks.