Sign In

Joshua Neil

Chief Data Scientist

Joshua Neil is a seasoned data scientist currently serving as the Chief Data Scientist at Securonix, a position he held from June 2021 until August 2023. In this role, he was instrumental in building and leading a team of engineers and data scientists, focusing on product innovation and improvement, which resulted in significant revenue impacts for the company.14 His expertise lies in leveraging machine learning and data-driven solutions to enhance cybersecurity measures.

Professional Background::

  • Securonix (Chief Data Scientist): Joshua led the data science team, focusing on developing advanced analytics for threat detection and response. He played a key role in designing strategies that contributed to multi-million dollar revenue impacts.12
  • Microsoft (Principal Data Science Manager): Prior to his tenure at Securonix, he managed the Microsoft Defender for Endpoint and Microsoft 365 Defender data science teams, where he led initiatives in insider threat detection and predictive analytics.13
  • Ernst & Young (Senior Manager): He developed cutting-edge security analytics solutions for global clients, significantly enhancing their cybersecurity posture.1

Education:: Joshua holds multiple degrees:

  • Ph.D. in Statistics from The University of New Mexico
  • MS in Electrical Engineering from the University of Southern California
  • BS in Mathematics from the University of California, Irvine, graduating Magna Cum Laude.1

Awards and Recognition:: He has received notable accolades, including the R&D 100 Award for his work on the PathScan network anomaly detection system, highlighting his contributions to innovation in cybersecurity.1

In addition to his professional roles, Joshua is also involved in academia as a member of the Industrial Review Board for the University of Washington's Masters of Data Science program, where he provides mentorship and strategic guidance.1

Highlights

Sep 5 · linkedin

It's important to understand what large language models are good at, and what they're not so good at.

One thing they're not so good at: handling non language data...duh...they're large LANGUAGE models.

Another thing they're not good at: handling structured data. Simply put, if we know the fields in the data, there are more accurate, less expensive ways to process them than LLMs. (The same could be said for any deep learning method, btw).

In cyber security, most of the data we deal with are structured, non language data. EDR logs, cloud logs, network telemetry all satisfy these conditions, and therefore there are better methods than llms to handle them.

Email contents are an exception, as are scripts (like powershell). These are broadly unstructured and language based, and I've seen good applications of deep learning to these data (not generative ai, but deep learning - the underlying approach to generative ai without the generative part).

But don't take it from me:

Mar 16 · linkedin

Is the rand() function generative AI? Some might say "well it doesn't learn from data". Fine, how about rand() * x, where x is some number learned from the data?

This would generate a sample uniformly distributed from 0 to x. Is this generative ai??

Related Questions

What are some of Joshua Neil's most notable achievements at Securonix?
How did Joshua Neil's work at Microsoft 365 Defender impact cybersecurity?
What strategies did Joshua Neil implement at Securonix for product improvement?
Can you provide more details about the PathScan network anomaly detection system invented by Joshua Neil?
What are the key differences between IOC-based detection and behavioral analysis according to Joshua Neil?
Joshua Neil
Joshua Neil, photo 1
Joshua Neil, photo 2
Add to my network

Location

Redmond, Washington, United States