Asks for

The ad says“Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).”
The ad says“Experience building or maintaining CLI tools in production environments.”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Strong proficiency in Go — the TokenShift CLI is written in Go, and this is where the core work happens.”
The ad says“Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).”
The ad says“Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).”
The ad says“Ability to work across platforms (macOS and Linux) and understand their differences at the system level.”
The ad says“Ability to work across platforms (macOS and Linux) and understand their differences at the system level.”
The ad says“Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”
The ad says“Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.”

About PointFive

PointFive finds the cost buried inside enterprise infrastructure, from the cloud to coding agents. We pioneered DeepWaste across cloud and data platforms, and we are now doing the same for AI token spend with TokenShift.
Fanatics, H&M, Hertz, Nubank, Citizens Bank and other Fortune 500 companies use PointFive to see and cut what their infrastructure and AI actually cost.

We have raised $96 million, most recently a $60 million Series B led by Accel. Engineering sits in Tel Aviv and our go-to-market team is focused on North America. PointFive Labs, our research arm, publishes original work on how infrastructure and AI actually get paid for.

CRN named us one of the 10 hottest cloud computing startups of 2026, Forbes Israel put us on its Next Billion Dollar Startups list, and Redpoint selected us for its InfraRed 100, its ranking of the private companies shaping the future of cloud infrastructure.

About the Role

TokenShift is PointFive's newest product — a lightweight CLI proxy that sits between AI coding agents and the shell or server, routing to different models for perfect model to effort fit, compressing command output to cut token spend and surface fleet-wide visibility. As an Senior software Engineer (Endpoint), you'll own the layer that makes TokenShift work: the per-agent harnesses, the compression rules and strategies, and the installation and hook mechanics that get us onto developer machines — reliably, invisibly, and at scale. This is a rare chance to build foundational infrastructure for AI developer tooling, at a company where your work directly shows up in the product used by engineering teams at companies like Expedia and EON.

What You'll Do

Own the installation and enrollment mechanics — the CLI flows, MDM scripts, and platform-specific packaging (Windows msi/exe, macOS pkg, Linux deb/rpm) that put TokenShift on developer machines at fleet scale.
Instrument and monitor endpoint behavior through the telemetry pipeline, using signals to improve coverage and catch regressions.
Build and maintain a robust lab enabling automatic research and development of new capabilities and platforms by AI agents.
Build and maintain agent harnesses — the per-agent adapters that hook into Claude Code, Cursor, Copilot CLI, Codex, and future agents to route network and shell commands through the TokenShift proxy.
Develop and extend compression rules and strategies for new command types (git, build tools, infra CLIs, search, and more), ensuring token reduction never comes at the cost of semantic accuracy.
Maintain correctness and resilience of the proxy layer — compression must be reversible on demand, hooks must never block a developer's command, and failures must be silent and transparent.
Collaborate with the research team to translate newly supported agent architectures into production-grade harnesses.

What We're Looking For

Must-have

Solid understanding of Windows and Unix fundamentals: process model, shell execution, stdin/stdout/stderr, file permissions, and IPC.
Deep familiarity with MDM deployment patterns (Jamf, Intune, JumpCloud, Ansible).
Strong proficiency in Go — the TokenShift CLI is written in Go, and this is where the core work happens.
Experience building or maintaining CLI tools in production environments.
Attention to correctness and edge cases — you think about what happens when things go wrong before they do.
Ability to work across platforms (macOS and Linux) and understand their differences at the system level.

Nice to have

TypeScript experience.
Hands-on experience with AI coding agents (Cursor, Claude Code, Copilot CLI, Codex) as a user or developer.
Experience with semantic compression, summarization, or output parsing.

Our Tech Stack

Go, TypeScript, Cloudflare, Snowflake

Why PointFive

Founders with a track record

Built and sold IntSights to Rapid7. Backed by top-tier investors with deep conviction in this category.

Category-defining product

DeepWaste detection and agentic remediation with clear, measurable ROI for both engineering and finance teams.

Early-stage leverage

Your work directly shapes the product, the culture, and the go-to-market playbook.

You'll build the infrastructure layer for AI developer tooling

The part that actually runs on every developer's machine, quietly making every agent session cheaper and more visible.