CareerPlanGet AI match score →

Ai Engineer

💼 Full-time🗓 2026-07-29

Core

Building core primitives for autonomous AI agents focused on accounting, including a model-agnostic harness, distributed cloud runtime, and multi-layer memory architecture.

Role type

Founding AI Infrastructure Engineer

Builds

Model-agnostic agent harness, distributed cloud runtime for concurrent agents, advanced harness systems, and multi-layer memory architecture

Domain

AI Infrastructure / Accounting

Deliverable

production ML models

Required skills

Distributed systems design, AI infrastructure development, correctness and determinism in complex systems, ambiguity tolerance, system architecture definition

Preferred skills

Container orchestration, sandboxing, ephemeral compute, LLM tool-use, agent frameworks, RAG systems, context window management, memory compression, retrieval at scale

Technologies

Container orchestration tools, ephemeral compute platforms, LLM frameworks, RAG systems

Responsibilities

Design and build model-agnostic agent harnesses with structured reasoning and safe code execution; Develop distributed cloud runtime for hundreds to thousands of isolated concurrent agents; Implement advanced harness systems for dynamic tool search and context management; Architect multi-layer memory systems for long-term semantic retrieval

Seniority

Founding, hands-on IC

Rewrite
## About the Role We're building AI agents that serve as the backbone for the world's businesses — starting with accounting, the hardest domain AI has ever touched. Every number has to be right. Every decision has to be traceable. Every agent has to survive legal scrutiny. We're not building demos. We're building infrastructure that sits beneath how real businesses run. We're looking for a founding engineer who wants to own the technical DNA of everything — not contribute to a roadmap, but define one. ## The Mission As our founding engineer, you'll build the core primitives that make autonomous AI agents reliable, inspectable, and scalable — across a model-agnostic harness, a cloud runtime for thousands of concurrent agents, and the memory and context systems that make them intelligent over time. ## What You'll Build - A model-agnostic agent harness — structured reasoning loops, tool calling, safe code execution, parallel orchestration. Simple. Deterministic. Inspectable. - A distributed cloud runtime for hundreds to thousands of fully isolated concurrent agents — container orchestration, ephemeral compute, filesystem virtualization, execution tracing - Advanced harness systems — dynamic tool search, decision-time prompting, context pruning, compression, and compaction - A multi-layer memory architecture — working memory, episodic task memory, long-term semantic retrieval, and compression over time ## What We're Looking For - Deep experience building distributed systems or AI infrastructure from scratch - Strong instincts for correctness, determinism, and debuggability in complex systems - Comfort with ambiguity — you've shipped things that had no playbook - Someone who's looked at their own work and thought "I don't know anyone else who could have built this" ## Good to Have - Experience with container orchestration, sandboxing, or ephemeral compute - Familiarity with LLM tool-use, agent frameworks, or RAG systems - Understanding of context window management, memory compression, or retrieval at scale ## Why Join - This is the only seat — not a headcount, not a queue. One person shapes the architecture of everything - You're solving the hardest AI trust problem: accounting, where a single wrong digit has legal consequences - If you can build agents that survive accounting, you've built agents that can survive anything - Bengaluru-based, early stage — the kind of work you'll look back on and know you built something that lasted
Sourced via wellfound · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.
Apply on Wellfound ↗