Founding AI Engineer

Overview

Founding AI Engineer to build the agentic layer for ProgressPlus, a manufacturing operations platform, and set the architecture, standards and evaluation discipline for every AI feature in the product.

What you'll do

• Architect the agent runtime: a server-side loop with durable runs, retries, scheduled and event-driven triggers, and human-in-the-loop steps
• Build the tool layer that turns the catalog of transaction-safe business operations into typed, tenant-scoped, permissioned AI tools for internal agents and external use through MCP
• Make evals part of how the team ships: fixture-based tests of every AI behaviour, wired into CI from day one
• Close the correction loop: capture every override, rejection and edit as structured signal so evals and retrieval improve from real use
• Build agent auditing into the platform's tamper-evident action log: every agent step logged, attributable and replayable
• Drive the event catalog: which business facts the product announces, from the office to the shop floor
• Grow the semantic grounding: a machine-readable model of the business so agents retrieve facts instead of inventing them
• Own the agent authority model: what an agent may do, its budget, and how much autonomy each workflow gets; instrument the runtime so every step is metered and traced
• Run what you ship: your agents in production are yours, including the incident when one misbehaves

What you'll need

• Experience taking multiple LLM-powered systems from prototype to production, at least one of them agentic (multi-step, tool-using, running unattended)
• Product instincts and a builder's disposition: translate a planner's or quoting manager's workflow into an agent architecture without a spec, then write the code, ship it, and talk to the people who use it
• Strong backend and systems engineering experience (5+ years)
• Experience building event-driven systems; can explain why the transactional outbox exists; thinks in contracts, idempotency and failure modes
• Treats evals and observability as part of the definition of done, and has wired LLM evaluation and observability tooling
• Treats AI security as an engineering problem: injection containment, retrieved content handled strictly as data, and hostile inputs in the eval suite
• Clear opinions on build versus buy for agent stack components
• Comfortable being first: setting standards, writing them down, and reviewing others' work against them
• Strong written communication; works with a team split across time zones

Nice to have

• Shipped something in manufacturing, ERP, supply chain or industrial operations, or genuinely curious about how factories run
• Document extraction pipelines: structured extraction from PDFs and email, OCR cascades, confidence-scored human review

Details

• Location: Bengaluru, India
• Works on a .NET and Next.js platform on Azure with a small team of full stack engineers, a lead, and AI and platform specialists
• Team is split across time zones
• Tech stack: C#, .NET 10, Rebus, SQL Server, Next.js, React, TypeScript, Azure, Bicep, GitHub Actions, MCP, Claude Code, Linear, GitHub