Applied AI Engineer - India

About this job

Dscout is growing its smart and driven team to build the most flexible and powerful user experience research platform used by top global brands in technology, healthcare, and finance. The company's tools help teams gain deep insights into human behavior to create better products. As an Applied AI Engineer, you will tackle the unique engineering challenges of AI-native products, where outputs are non-deterministic and success means reliable agent behavior in real-world scenarios. A normal day involves owning the production improvement loop, instrumenting agent workflows, analyzing why agents underperform, designing targeted behavior upgrades, building backend services and APIs, and running controlled experiments or staged rollouts to measure business impact.

Responsibilities

• Own the production improvement loop across agent behavior, feedback, and experimentation.
• Instrument agent workflows to track model interactions, tool use, decisions, and failures.
• Define quality standards, evaluation datasets, regression coverage, and production monitoring.
• Investigate agent underperformance across context, instructions, tools, and guardrails.
• Design and ship behavior improvements including prompt updates, decision logic, and tool use.
• Build backend services, APIs, and feedback pipelines for observable and reproducible agents.
• Run controlled experiments and staged rollouts to measure quality and business results.
• Partner with Product, Data Science, and Sales to prioritize problems and ensure safeguards.

Requirements

• 2 to 5 years of software engineering experience with LLM-powered features or agents in production.
• Fluency with prompting and context engineering as an engineering discipline.
• Experience building or maintaining evaluation harnesses for AI systems and regression detection.
• Comfort with non-determinism and reasoning about agent behavior across production traffic.
• Experience running experiments like A/B tests or staged rollouts to validate changes.
• Track record of shipping features that real users depend on and owning post-launch results.
• High-agency mindset to independently investigate ambiguous underperformance issues.
• Comfort using AI coding tools like Cursor, Claude Code, or Copilot in daily workflows.

Good to have

• Experience with voice or real-time conversational AI systems.
• Familiarity with LLM observability and tracing tools such as LangSmith or Datadog.
• Experience with agentic orchestration frameworks like LangChain or LangGraph.
• Exposure to MCP-based tooling or agentic data workflows.

Who should apply

This role suits software engineers with two to five years of experience who excel at building, evaluating, and operating production-ready LLM systems and agentic workflows. Candidates should be comfortable handling AI non-determinism, eager to run rigorous experiments, and capable of investigating ambiguous performance problems. Freshers are not suited for this position as specific hands-on production AI engineering experience is required.

Frequently asked questions