Kushagra Sharma's Blog Posts

Kushagra Sharma

Kushagra Sharma is a Product Manager at Cisco working on AIOps products that help teams operate complex systems with intelligence and scale. He focuses on translating customer problems into practical, data-driven solutions at the intersection of AI, reliability, and software platforms. Passionate about clear thinking and strong product craft, Kushagra writes to share insights on technology, product strategy, and building systems that actually work in the real world.

Beyond Open Weights: From Model Choice to AI Leadership
Artificial Intelligence
4 Minute Read

Beyond Open Weights: From Model Choice to AI Leadership

Why model choice must lead to portable intelligence and governed action
AI Tokenomics: From Cost Visibility to Cost Discipline
Leadership
5 Minute Read

AI Tokenomics: From Cost Visibility to Cost Discipline

Dashboards show where tokens went, but sustainable AI economics starts with deciding which tokens should be spent at all.
The Data Abundance Paradox
Customers
8 Minute Read

The Data Abundance Paradox

Organizations are generating more machine data than ever before, yet only a fraction of that data drives measurable business outcomes.
Phantom Stealer Unmasked: Shellcode, Steganography, and Credential Theft
Security
15 Minute Read

Phantom Stealer Unmasked: Shellcode, Steganography, and Credential Theft

Unmask Phantom Stealer, a .NET credential-stealing threat, and explore its steganography, shellcode injection, and evasion tactics with 32 actionable Splunk detections.
The Tokenomics of Trust: Evaluate Everything, Not a Sample
Artificial Intelligence
7 Minute Read

The Tokenomics of Trust: Evaluate Everything, Not a Sample

Most teams evaluate 10% of AI traffic because judging quality is expensive, so the costly failures hide in the unchecked 90%. Here's how to score all of it.
When You Repatriate AI, the Invoice Disappears – The Cost Doesn’t
Artificial Intelligence
5 Minute Read

When You Repatriate AI, the Invoice Disappears – The Cost Doesn’t

Here's how to measure GPU, memory, and token cost together so you know what your AI really costs.