AI Engineer, Harness & Eval

NAVEX; Bengaluru; Hybrid; INR 2,700,000 - 3,300,000; Full time

Benefits offered by this job

Hybrid work environment

Competitive compensation

Career growth opportunities

Job summary

NAVEX in Bengaluru is seeking an experienced AI engineer to join its AI Centre of Excellence. You will build production agents, design eval harnesses, and ensure safe, compliant operation.

You'll own end-to-end systems, deploy on cloud, and work in a fast-paced hybrid environment that values speed and measurable outcomes. This role emphasizes collaboration, writing, and contributing to enterprise-grade AI for business operations.

Qualifications

• 3+ years of AI engineering experience building LLМ systems used by real users.
• Hands-on with agentic harnesses and eval frameworks.
• Experience deploying on major clouds (AWS, GCP, or Azure).

Responsibilities

• Build and ship agents for business operations and internal products.
• Design the eval harness and testing pipelines.
• Instrument production agents with tracing and latency metrics.
• Integrate MCP servers and safe data access.
• Take over externally built agents and improve observability.
• Ensure compliance and verification before release.

Skills

Python

LLM systems

Agent harnesses

Cloud deployment

AWS

Tools

Claude Code

Codex

Job description

You'll be one of the first engineers in NAVEX's AI Centre of Excellence, a new function with a company-wide mandate to put AI into how the business runs. We build AI for business operations and internal agentic products.

You'll build production agents and, more importantly, the harness around them: the evals, tracing and guardrails that tell us an agent works before a user finds out it doesn't. Some agents you'll build from scratch. Others were built by external partners and are moving in-house. You'll own both.

At NAVEX, we’re transforming the world-making it safer, more ethical, and ensuring every voice is heard. That’s real impact.

Our high-performance culture is driven by our values. We move with speed, passion and purpose - as one team. We are bold in our ideas, accountable in our actions, and committed to doing the right things right.

You’ll thrive in this hybrid role surrounded by an engaged, collaborative team deeply committed to innovation, speed, and measurable outcomes. Join us and help shape what’s next!

What you’ll get:

• Meaningful Purpose. Your work helps organizations operate with integrity and protect their people-at a scale few companies can match.
• High-Performance Environment. We move with urgency, set ambitious goals, and expect excellence. You’ll be trusted with real ownership and supported to do the best work of your career.
• Candid, Supportive Culture. We communicate openly, challenge ideas- not people-and value teammates who embrace bold thinking and continuous improvement.
• Growth That Matters. You can count on authentic feedback, strong accountability, and leaders invested in your success so you can achieve real growth.
• Rewards for Results. We provide clear, competitive compensation designed to recognize measurable outcomes and real impact.

What you’ll do:

• Build and ship agents for business operations and internal products. End to end: retrieval, tool use, orchestration, state, and handoffs to human
• Build MCP servers and tool integrations that give agents safe, reliable access to company systems and data
• Design the eval harness. Golden sets, regression suites, LLM-as-judge where it earns its place, human review loops. Every change ships with evidence that it made things better
• Instrument agents in production. Tracing, failure taxonomies, cost and latency tracking. When something degrades, you find out first and know why
• Take over agents built by external partners. Read the code, find where it breaks, add the evals and observability it shipped without
• Hold a verification bar. In a compliance company, an answer we can't source or explain doesn't ship

What you’ll bring:

• 3+ years of AI engineering experience, most of it building LLM systems that real users depend on
• Hands-on experience building with agentic harnesses such as Claude Code, Codex, or similar
• You've built MCP servers, eval harnesses or LLM Systems, not just used them and can explain how you knew a change was an improvement
• You've owned a system that degraded in production and diagnosed why, and can walk us through it
• Strong Python. Comfortable with at least one agent framework, and willing to go below it when it gets in the way
• Experience deploying on a major cloud provider: AWS, GCP or Azure
• You've shipped end-to-end and own outcomes, not just tasks
• You can take a non-technical team's workflow and turn it into a working system
• Building RAG over messy enterprise documents—proposals, contracts and policies
• Working with AWS Bedrock or inside VPC-constrained enterprise environments
• Building in regulated or compliance-heavy domains
• Sharing your work through writing, talks or open source
• Culture Agility. Comfort working in a fast-paced, candid environment that values innovation, healthy debate, and follow-through
• Fuel performance and outcomes. Leverage your job competencies and champion NAVEX’s core values

Our side of the deal:

• We’ll be clear, we’ll move fast, and we’ll invest in your success. You deserve to be supported, challenged, and rewarded for the impact you make—and we commit to doing that every step of the way.
• The starting pay for this role is 3,000,000 INR with 10% MBO. Discover how you can grow, lead, and make an impact by visiting our career page to learn more. NAVEX is an equal opportunity employer committed to including individuals of all backgrounds, including those with disabilities and veteran status.