Medium

Prashant Kumar

In

Data-Agents Dojo

by

Prashant Kumar

·

Nov 2

Evals: The Inverted Ownership Problem in Enterprise AI

Who controls model change should determine how enterprises evaluate their AI systems

Evals: The Inverted Ownership Problem in Enterprise AI

In

Data-Agents Dojo

by

Prashant Kumar

·

Oct 30

Introducing Hex Agent OS: A Message-Bus Runtime for Enterprise Multi-Agent Systems

This article introduces Hex Agent OS, its architecture, the design principles behind it, how it is designed to scale, and which parts are…

Introducing Hex Agent OS: A Message-Bus Runtime for Enterprise Multi-Agent Systems

In

Data-Agents Dojo

by

Prashant Kumar

·

Jul 20

The Telco Context Graph, Part 4: Where to Start

Part 1 made the case. Part 2 went into the architecture. Part 3 defined the hard limits. This piece is about the first version — what a…

The Telco Context Graph, Part 4: Where to Start

In

What AI Actually Does

by

Prashant Kumar

·

Jun 3

LLMs are not magic — they’re statistics

LLMs are not magic — they’re statistics

LLMs are not magic — they’re statistics

In

Data-Agents Dojo

by

Prashant Kumar

·

Jun 3

The Challenges of Building and Delivering “Talk to Your Data” — AI Analyst & Data Scientist

The pitch was seductive. Natural language over SQL. Remove the technical barrier and business users could finally engage with their data…

The Challenges of Building and Delivering “Talk to Your Data” — AI Analyst & Data Scientist

In

Data-Agents Dojo

by

Prashant Kumar

·

May 19

The Telco Context Graph, Part 3: The Hard Limits

Parts 1 and 2 built the case for autonomous network operations and described the architecture that enables it. This piece covers the…

The Telco Context Graph, Part 3: The Hard Limits

In

Data-Agents Dojo

by

Prashant Kumar

·

Apr 29

The Context Graph for Telco Network Intelligence: What the Five Layers Actually Require — Part 2

Part 1 made the case for the context graph in telco — four use cases, a blueprint, a direction. This piece is the layer beneath that: what…

The Context Graph for Telco Network Intelligence: What the Five Layers Actually Require — Part 2

In

Data-Agents Dojo

by

Prashant Kumar

·

Apr 13

The Case for Agentic AI Investment: A Practical Framework for Network Operations

A staging model for engineering teams to build credible, risk-calibrated business cases for agentic AI — from first pilot to autonomous…

The Case for Agentic AI Investment: A Practical Framework for Network Operations

In

Data-Agents Dojo

by

Prashant Kumar

·

Apr 1

The Token Trap: Provider Cost vs. Buyer Value

The AI industry chose tokens as its billing unit for perfectly good infrastructure reasons. Then enterprises made the mistake of using them…

The Token Trap: Provider Cost vs. Buyer Value

In

Data-Agents Dojo

by

Prashant Kumar

·

Mar 24

What Are Agents Actually Running On? The Case for Metadata as the New Oil

The bottleneck in enterprise AI isn’t the model. It’s the metadata.

What Are Agents Actually Running On? The Case for Metadata as the New Oil

Prashant Kumar

Prashant Kumar

Building enterprise AI agents | Cardinality founder (exited) | AI Analyst, Data Scientist on System of Record agents, semantic operations & agentic systems

Following

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech