San Francisco Bay Area
I build products that use machine learning and artificial intelligence to solve a variety of internet scale network security problems. Some of this work was presented at BlackHat, DefCon and FIRST. During my time at OpenDNS and Cisco I've built a successful track record of working closely with customers to translate their security requirements into proven security products. Products we've brought to market solved difficult problems arising in varied areas like malware detection, threat intelligence, and large scale network configuration. Products like the malware detection suite impact millions of endpoint devices and provide critical ransomware detection capabilities. The solutions integrated classical machine learning methods with generative AI techniques leading to novel products in the space.
Helped design and build the first AI Assistant for network management. Cisco Secure AI Assistant helps you create network policies, troubleshoot network issues, and optimize your security rules. The Secure AI Assistant helps bring down the average rule creation time and average time for network troubleshooting. Integrated within the networks of multiple F500 companies. Helped drive increased customer retention and growth for the cloud business.
- Helped design and build Country IP Risk feature for Cisco Umbrella. This feature was requested by multiple national telecos. Country IP Risk allowed telecos to understand the threats facing their national IP space. We focused on identifying IP ranges most susceptible to attack types associated with large monetary damages. These new features within Cisco Umbrella helped grow our business amongst European telecos. - Built a new ransomware and exploit kit detection product for Cisco Umbrella. Customers required new detection methods to identify potential ransomware domains. I worked with customers to understand their requirements and then led the design of a new time series classification system. System ran over daily global DNS data logs (approximately 30 TB a day) and led to two patents. The new system led to Umbrella winning deals in the MSP and SMB spaces. - Helped build out Cisco Umbrella's Covert Communication detection product line. The product helped users identify covert channels like DNS tunneling in their network. Deployed across all 27 worldwide resolvers. This product solidified our detection product suite and led to greater growth in the healthcare sector. Work led to a patent and talk at FIRST 2019. - Designed database to store SSL certificates from internet wide scan. This database was used to track botnet actors. Work was discussed in BlackHat 2016 talk. - Used graphical methods to identify potentially compromised home routers based on global SSL certificate data (DefCon 2017)
Led the development of a malware spam system which identified IPs and domains associated with phishing and mass spam emails. I developed this system after talking with prospective clients in the oil and gas industries. The system ran over our daily global authoritative DNS data (3 TB) and gave clients a prioritized list of IPs and domains to block.
Worked with the research team to develop and deploy machine learning models to detect DNS and other network related threats.