Principal Engineer

Sage City; Bengaluru; Hybrid; INR 1,800,000 - 2,500,000; Full time

Job summary

Sage, Bengaluru, seeks an AI Infrastructure Engineer (Mid/Senior) to design and operate scalable agent platforms and AI tooling. You will orchestrate tools, registries, and sandboxed execution, while integrating with CI/CD and codebases across teams.

You will lead memory management, evaluation pipelines, and feedback loops to continuously improve agent performance and reliability, in a hybrid work setup.

Qualifications

• Requires 3+ years in software engineering (mid) or 6+ years (senior).
• Experience building production systems in cloud environments (AWS/GCP/Azure).
• Hands-on with containers, Kubernetes, and infrastructure as code.
• Direct work with LLMs: tool use, agents, prompt/context engineering, RAG and evaluation.
• Familiar with agent protocols and ecosystems; experience with AI coding tools.
• Strong communication to explain trade-offs to technical and non-technical audiences.

Responsibilities

• Build and operate agent platforms: orchestration, tool registries, plugin frameworks, sandboxed execution.
• Develop coding and review agents that work on large codebases and integrate into SDLC.
• Design context/memory management: retrieval, summarization, session state, long-term memory.
• Create evaluation frameworks (offline/online) to measure agent improvements.
• Create feedback loops from traces to prompts, tools, skills, and evals.
• Develop internal SDKs and developer tooling for cross-team adoption.
• Own observability for agent systems: tracing, cost, latency, quality monitoring.
• Collaborate with engineers to identify high-value workflows and iterate on usage data.

Skills

System design

API design

Data modelling

Cloud platforms

Containers & Kubernetes

Infrastructure as code

Distributed systems

LLMs & agent tooling

Prompt engineering

RAG & memory management

Project leadership

Mentoring

Communication

Observability & CI/CD

Tools

Kubernetes

Docker

CI/CD tooling

Job description

Job Description AI Infrastructure Engineer (Mid / Senior)Location: Bengaluru (Hybrid – 3 days a week in office)About UsSage is a global leader in accounting and business management software, trusted by millions of businesses worldwide. Together with our Bengaluru engineering team, we are building finance, accounting, payroll and expense management solutions that simplifies how businesses track and reconcile spending.Thousands of finance teams globally use our product to streamline expense management, integrate seamlessly with leading accounting platforms, and are recognized for its clean design, usability, and speed.We're investing heavily in AI — building the platform that lets AI agents and assistants do real engineering work, reliably and at scale. This team designs and runs the infrastructure behind our coding agents, review agents, and internal AI assistants: the orchestration, context and memory systems, evaluation frameworks, and feedback loops that make them better over time.What We're Looking ForWe're looking for someone with 3+ years of software engineering experience (mid-level) or 6+ years (senior), with hands-on experience building LLM-powered or agent-based systems in production.Technical Skills Solid software engineering fundamentals: system design, API design, data modelling, and clean, maintainable code. Experience building and running production systems in the cloud (AWS, GCP, or Azure), including containers, Kubernetes or similar, and infrastructure as code. Hands-on experience with distributed systems: queues, async workflows, concurrency, reliability, and failure handling. Practical experience building with LLMs: tool use, agents, prompt and context engineering, RAG, and evaluation. Familiarity with agent protocols and ecosystems (e.g., MCP) and with using AI coding tools day to day.Leadership & Ownership Comfort working independently and handling ambiguity: you can take a loosely defined problem, scope it, and drive it to production. Clear written and verbal communication, including explaining trade-offs to both technical and non technical audiences. At the senior level, we also expect you to lead projects and set technical direction, mentor other engineers, define evaluation and quality standards, and influence architecture across teams.Mindset & Collaboration Growth mindset — curious, always learning, and open to feedback. Works closely with the engineers who use these tools every day; you help the whole organization build software better and faster.Nice to Have Experience with trace analysis or log mining to find and fix agent failure modes. Experience building internal platforms or SDKs that other teams adopted. Background in observability, CI/CD, or developer productivity tooling. Experience with model routing, cost optimization, or prompt caching. Open-source or public contributions in agent tooling or LLM infrastructure.What to Expect We value humility and collaboration — we're looking for people who can work as part of a team, not in silos. We expect engineers to take ownership beyond just writing code — including tests, documentation, and long-term system quality.#LI-RTKey Responsibilities The Impact You'll Make Build and operate agent platforms: orchestration, tool and skill registries, plugin frameworks, and sandboxed execution. Develop coding agents and review agents that work well on real, large codebases, and integrate them into the existing SDLC (IDE, CI/CD, code review, issue tracking). Design context and memory management: retrieval (RAG), summarization and compression, session state, long-term memory, and repository-aware context. Build evaluation frameworks — both offline (benchmarks, regression suites) and online (production quality signals) — so we can measure whether agents are improving. Create feedback loops: capture and mine traces, find failure patterns, and turn them into better prompts, tools, skills, and evals. Build developer tooling and internal SDKs that make it easy for other teams to build on and adopt the platform. Own observability for agent systems: tracing, cost and latency tracking, and quality monitoring. Work with engineers and stakeholders to identify high-value workflows, ship quickly, and iterate on real usage data.#LI-RT