Research Engineer - Agent Intelligence & Evaluation

Ixigo; Gurugram District; On-site; INR 2,600,000 - 4,000,000; Full time

Job summary

Ixigo in Gurugram is seeking an ML engineer to own the intelligence layer of self-healing voice agents for enterprise support. You will design evaluation metrics for voice and ASR, build tracing across audio, STT, LLM reasoning and tool usage, and drive feedback loops to improve agent reliability on real traffic.

Candidates should have 3–5 years in ML engineering or research roles, strong Python skills, and deep experience in at least two areas such as speech models, LLM agents, or eval and

Qualifications

• 3 to 5 years in ML engineering, research engineering, or applied research.
• Strong Python and modern ML tooling.
• Depth in at least two: speech and audio models, LLM agent systems, and eval or observability infrastructure.

Responsibilities

• Evaluation infrastructure for audio-native metrics and adversarial datasets.
• Observability across the pipeline linking audio, STT, LLM reasoning and tool calls.
• Self-improvement systems: mine production traces, generate training data, validate fixes, guard against regressions.

Skills

Python

ML tooling

Speech models

LLM agents

Observability infra

Job description

Job Summary

Voice agents fail in ways traditional software doesnt. ASR confidence drops on an accent and a tool call misfires. Latency breaks turn-taking and the LLM hallucinates a policy. A model swap silently regresses production and nobody catches it for a week.

Were building self-healing voice agents for enterprise customer support. This role owns the intelligence layer: the evals that catch failures before shipping, the observability that traces them across the pipeline, and the feedback loops that let agents fix themselves

What youll own

• Evaluation infrastructure. Audio-native metrics for barge-in, prosody, and turn-taking. Adversarial datasets across accents and edge cases. LLM-as-judge rubrics for task success, tool-use correctness, and recovery.
• Observability across the pipeline. Tracing that correlates audio, STT, LLM reasoning, tool calls, and TTS to a single conversation. Analysis and alerting that surfaces cascade failures instead of hiding them.
• Self-improvement systems. Mine production traces for failure patterns, generate targeted training or prompt data, validate fixes with adversarial replay, and guardrail against regressions.

Qualifications

• Who we're looking for - 3 to 5 years in ML engineering, research engineering, or applied research. Strong Python and modern ML tooling. Depth in at least two of: speech and audio models, LLM agent systems, and eval or observability infrastructure.
• Youve shipped something non-trivial where research met production. You read papers, spot when a benchmark measures the wrong thing, and translate ideas from Interspeech, ACL, or NeurIPS into systems that run on real traffic. Publications welcome, not required.
• Nice to have Real-time systems or telephony experience. Work on RLHF, DPO, or synthetic data pipelines. Familiarity with enterprise deployment (SOC 2, PII, data residency).