Agent Evaluation & Instrumentation Engineer

Location: Pune | Experience: 8–11 Years | CTC: Up to ₹33 LPA(576)
Notice Period: Immediate–45 Days | Preference: Local Pune Candidates
Interview: 2 Technical Rounds + Client Round + HR

MANDATE CRITERIA – NON-NEGOTIABLE

- 6+ years in ML/Data/Software Engineering with robust evaluation/quality focus

- Hands-on LLM/ML evaluation experience

- Strong Python skills

- Experience with Promptfoo / DeepEval / custom evaluation frameworks

- Strong understanding of LLM/Agent evaluation design & statistical rigour

- Hands-on data analysis & metric interpretation

- Experience with tracing & observability tools

- Strong technical documentation, analytical & stakeholder-management skills

- Understanding of Telco customer intents & customer journeys

- 2+ years stability in an organization

KEY RESPONSIBILITIES

- Design and maintain LLM/Agent evaluation suites, golden sets, regression packs & adversarial tests





- Build continuous evaluation & scoring pipelines

- Conduct quality/operability gate reviews and reproduce evaluation results

- Analyse model regression, failure modes, defects & evaluation trends

- Calibrate quality thresholds, judges & evaluation datasets

- Monitor evaluation drift and maintain hold-out/golden sets

- Prepare technical quality reports & evaluation findings

- Collaborate with AI/ML, Engineering & Operations teams on quality improvements

- Mentor team members and communicate findings to technical/client stakeholders

⭐ GOOD TO HAVE

Agentic AI | RAG | Telco AI | AI Red-Teaming | Responsible AI & Safety | Judge Calibration | Management Presentations

Ideal Profile: AI/ML Evaluation + Python + Data Analysis + LLM/Agent Quality with strong communication and client-facing skills.

📌 Agent Evaluation & Instrumentation Engineer (Pune)
🏢 Umanist Staffing
📍 Pune