20 AI Projects Shipped
10 Years in Cybersecurity/Observability
$4.3B M&A Exposure • $500M Impact
FEATURED PROJECTS
Built and shipped a 4-agent autonomous system on real SOC telemetry
Results:
MTTR: 60 → 3 minutes
Accuracy: 60% → 98%
166% ROI (Forrester)
Featured at AWS re:Invent
Built business-impact-driven AutoML platform optimizing for ROI, not accuracy
Results:
Beat IBM / SAP / Microsoft in competitive deals
First product sold live at Gartner Summit
$100M+ valuation in 90 days
Reframed failed AI product into asset-life optimization platform used in production
Results:
$42M pipeline from single event
Became the flagship Baker Hughes platform
Running 7+ years later
20 AI PROJECTS SHIPPED
-
Built and shipped a 4-agent autonomous investigation system on real SOC telemetry
Reduced MTTR from 60 minutes to under 3 minutes
Increased investigation accuracy from 60% → 98%
Featured at AWS re:Invent • 166% ROI (Forrester)
-
Built context-aware query assistant preventing costly over-broad queries
Eliminated “blank start” problem with continuous suggestions
Reduced compute waste, saving ~$2.5M annually
-
Built multi-turn natural language interface for security queries
Reduced query syntax errors by 78%
Shipped to production serving thousands of queries daily
-
Built agentic system converting analyst intent into production-ready detection rules
Automated rule creation, translation, and maintenance
Eliminated manual syntax authoring
-
Built conversational investigation interface directly in Slack
Enabled SOC analysts to query and investigate without leaving workflow
Increased investigation speed and adoption
-
Led AI product evaluation and integration for Dexda, Unomaly, and Airbrake
Unified disparate ML systems into a single production platform
Contributed to $50M → $200M ARR growth
-
Integrated Unomaly anomaly detection into observability platform
Automated detection of unusual log patterns across large-scale systems
Improved signal detection without manual tuning
-
Integrated Dexda predictive fault detection to suppress alert storms
Reduced alert fatigue and improved operator efficiency
Enabled scalable alert correlation across environments
-
Built ML-driven system to automatically adjust alert thresholds
Eliminated manual configuration and tuning
Improved accuracy and reduced false positives
-
Built business-impact-driven AutoML platform optimizing for ROI, not accuracy
Shipped from concept to GA in 90 days via rapid iteration
First product sold live at Gartner Summit
$100M+ valuation within 3 months
-
Reframed failed AI product into asset-life optimization platform
Generated $42M pipeline from a single event
Became flagship Baker Hughes product still in use today
-
Repositioned AI from “replacement” to “decision support”
Improved adoption by aligning with operator workflows
Reduced resistance and increased deployment success
-
Built intelligent alert correlation and escalation system
Reduced noise while surfacing critical events
Deployed in BP flagship industrial environment
-
Built internal observability platform validating acquisition decision
Demonstrated product-market fit prior to acquisition
Supported executive M&A strategy discussions
-
Built conversational system for security rule creation
Enabled natural language → detection logic
Early precursor to modern AI copilots
-
Built Slack-based system for querying security data
Enabled analysts to investigate without switching tools
Early agentic workflow prototype
“Greg turns simple ideas into state-of-the-art products — bridging users, design, and engineering.”
— Tejaswi Redkar
CEO & Founder (former Cisco, AppDynamics, Sumo)
“Greg shipped a production-ready multi-agent system that set Sumo’s AI direction.”
— Brandon Borodach
Field CTO, Abstract Security
“Greg validates AI product-market fit with real customers — not assumptions.”
— Catherine Davis
VP Product Management, Addigy
“Greg improved multi-agent AI reliability by elevating RAG quality and system design.”
— Steve Berube
Technology Leader, Sumo Logic