---
type: "post"
id: "cisco-president-jeetu-patel-says-a-27-year-a7672e"
slug: "cisco-president-jeetu-patel-says-a-27-year-a7672e"
title: "Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to"
description: "Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to refactor most of Cisco’s legacy code."
retrieval_nugget: "Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to refactor most of Cisco’s legacy code. The stated target is 98% in 24 months, with a four-week proof period. It is a notable modernization bet, but the report describes a goal, not a demonstrated result."
published_at: "2026-11-03"
updated_at: "2026-11-03"
record_date: "2026-11-03"
date_kind: "published_at"
topics: ["coding-agents","ai-adoption","future-of-work"]
entities: ["Cisco","Jeetu Patel","Padmanabhan Krishnamurthy","EzDubs"]
related_patterns: []
novelty_label: "new_fact"
what_changed: "Cisco’s Jeetu Patel gave 27-year-old Padmanabhan Krishnamurthy unlimited AI tokens to modernize 98% of Cisco’s code; the priors cover Uber’s code cleanup and general AI modernization, not this Cisco project."
prior_refs: ["https://newruntime.com/posts/uber-uses-a-similar-setup-to-clean-up-200090/"]
source_url: "https://finance.yahoo.com/technology/articles/cisco-chief-found-27-old-070000750.html"
source_urls: ["https://finance.yahoo.com/technology/articles/cisco-chief-found-27-old-070000750.html"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:cisco-president-jeetu-patel-says-a-27-year-a7672e"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/cisco-president-jeetu-patel-says-a-27-year-a7672e/","markdown":"https://newruntime.com/posts/cisco-president-jeetu-patel-says-a-27-year-a7672e.md","json":"https://newruntime.com/posts/cisco-president-jeetu-patel-says-a-27-year-a7672e.json"}
markdown_url: "https://newruntime.com/posts/cisco-president-jeetu-patel-says-a-27-year-a7672e.md"
---

# Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to

## Retrieval answer

Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to refactor most of Cisco’s legacy code. The stated target is 98% in 24 months, with a four-week proof period. It is a notable modernization bet, but the report describes a goal, not a demonstrated result.

## Summary
Cisco president Jeetu Patel says a 27-year-old engineer is testing whether a two-person team can use AI to refactor most of Cisco’s legacy code. The stated target is 98% in 24 months, with a four-week proof period. It is a notable modernization bet, but the report describes a goal, not a demonstrated result.

## What happened
Fortune reports that Cisco president and chief product officer Jeetu Patel identified Padmanabhan Krishnamurthy, known as “Paddy,” as a candidate to take on a large legacy-code project. Krishnamurthy is 27 and joined Cisco in 2025 after the company acquired EzDubs, the real-time translation startup he co-founded in 2023. Patel said Krishnamurthy had been working in an area he considered unstrategic before the two spoke at Cisco’s San Francisco office.

In that conversation, Krishnamurthy said he believed he could rewrite Cisco’s legacy code with a two-person team. Patel told Fortune he had given him “unlimited tokens” for refactoring and told him to proceed. The article does not name the model, coding assistant, token provider, repository, programming languages, or the specific systems included in the work. “Unlimited” is Patel’s description of the allocation, not a published usage policy or cost figure.

The initial test is time-bounded: Patel says he gave the project four weeks to prove itself. The broader stated objective is to refactor 98% of Cisco’s legacy code within 24 months. The source does not say whether the four-week proof period has concluded, what deliverable would count as proof, or whether Cisco has formally approved the 24-month target as a company-wide program. No code-quality, defect, security, throughput, or cost results are reported.

The initiative also figures in Patel’s view of how Cisco should organize teams. He says he wants early-career employees to challenge assumptions held by more experienced staff, and acknowledges that this can create friction. The article describes him seeking more employees with Krishnamurthy’s willingness to question established practices. It also recounts a separate intern example, but provides no technical details sufficient to connect that work to the legacy-code effort.

## What the archive adds
[Uber uses a similar setup to clean up old code.](https://newruntime.com/posts/uber-uses-a-similar-setup-to-clean-up-200090/) covered Uber’s system for deleting stale feature-flag code, including AutoClean, PolyglotPiranha and MCP. Cisco’s reported plan is a new company initiative on the same broad problem of using agents to change legacy software; it is not evidence that Cisco uses Uber’s tools or approach.

## Trend signal
The supplied archive contains no directly applicable linked trend record beyond the related legacy-code case.

## Why now
The article frames the initiative as a response to a direct conversation between Patel and Krishnamurthy, rather than a previously announced program. Patel says he wants a four-week proof before pursuing the 24-month ambition. That sequencing makes the near-term test important: it could clarify whether the proposed team and approach can produce results, but no outcome is yet reported.

## Connections to other events
The Cisco initiative meets the legacy-code cleanup problem described in the earlier New Runtime note on [Uber’s stale feature-flag deletion system](https://newruntime.com/posts/uber-uses-a-similar-setup-to-clean-up-200090/). Uber’s case centers on identifying and removing obsolete feature-flag code; Cisco’s stated target is much broader, but the source does not establish that its scope or implementation is comparable.

## Why it matters
Legacy modernization is often constrained less by the ability to produce code than by the risk of changing systems whose dependencies and behavior may be poorly understood. Cisco’s reported ambition is striking because it ties an unusually broad target to a small team and a short initial proof period. If that experiment succeeds, it could encourage other large organizations to test AI-assisted refactoring against tightly scoped milestones rather than treating code generation as an end in itself.

The useful lesson for engineering teams is not to copy the 98% figure. Cisco has not published what counts as legacy code, how it will measure completion, or what safeguards it will require. Teams considering similar work should first define bounded modules, acceptance tests, rollback paths, and human ownership of changes. Without those details, speed or volume alone would not show that a refactor is safe or maintainable.

## Limits and uncertainty
The account is based on Patel’s remarks to Fortune and does not include a technical briefing from Cisco, Krishnamurthy, or the project team. It provides no baseline for the codebase, definition of “rewrite” or “refactor,” or evidence that AI-generated changes have been merged. It also omits the tools, model, token costs, review process, test coverage, and risk controls. The 98% and 24-month figures should therefore be read as a stated goal, not a measured forecast or completed outcome.

## Skill to build now
For one low-risk legacy module, define a refactoring proof before asking an AI coding tool to change it: record its existing behavior in tests, specify the allowed scope, and require a human-reviewed diff plus passing regression checks. Track defects and review time alongside lines changed. This makes a short experiment informative without assuming that generated code is automatically a successful modernization.

## Reader takeaway
Cisco’s report is a high-stakes proposal, not proof that AI can safely rewrite nearly an entire enterprise codebase. The important test is whether the four-week experiment can produce verifiable improvements that justify the 24-month target. Until Cisco shares its method and results, teams should treat the numbers as an ambition and focus on measurable, reviewable slices of legacy code.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Coding Agents](https://newruntime.com/topics/coding-agents.md): Explore the coding-agents topic hub.
- topic hub: [Ai Adoption](https://newruntime.com/topics/ai-adoption.md): Explore the ai-adoption topic hub.
- topic hub: [Future Of Work](https://newruntime.com/topics/future-of-work.md): Explore the future-of-work topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)