Website:
kmmtech.ai
Job details:
We are looking for Offshore role- Agentic AI Engineer
It will be a Full-time employment with KMM technologies.
Please connect to me to discuss for the below opportunity with KMM Technologies.
Offshore/India- 100% remote
Agentic AI Engineer
Work Schedule [ 5.30 PM – 3.30 AM IST]
Should join in 2-3 weeks.
KMM is investing in an “Agentic AI Platform for Observability and Cyber Vulnerability Management”.
About the Team & Role
We are looking for an AI Engineer for Observability to build the next-generation monitoring and evaluation "nervous system" for our AI-powered products and internal infrastructure.
You will sit at the intersection of Site Reliability Engineering (SRE), MLOps, and AI/LLM application development. While traditional observability focuses on CPU, memory, and HTTP status codes, your focus will be on tracking non-deterministic AI behavior: prompt drift, token cost optimization, retrieval-augmented generation (RAG) faithfulness, vector DB latencies, and agentic step-by-step trace analysis.
Key Responsibilities
- Build AI-Native Observability Pipelines: Design and scale telemetry architectures (OpenTelemetry, Prometheus, distributed tracing) capable of capturing LLM inputs, outputs, spans, and tool calls at production scale.
- Implement Evaluation & Guardrails: Build and maintain automated evaluation loops, regression tests, and real-time guardrails to catch hallucinations, bias, and schema failures before they reach end-users.
- Develop AI-Driven Remediation & Dashboards: Create automated root-cause analysis tools, AI-assisted alerting workflows, and custom dashboards that help engineering and SRE teams debug complex distributed and AI workloads.
- FinOps & Resource Tracking: Implement cost-attribution models and token-usage monitoring to track, bound, and optimize GPU, model-serving (vLLM/Triton), and third-party API spend per tenant or feature.
- Cross-Functional Collaboration: Partner with ML engineers, data engineers, and product teams to define SLIs/SLOs for AI features and establish standards for model logging and tracing.
Required Skills
- Experience: 5+ years in software engineering, platform engineering, or SRE, with at least 2+ years dedicated to AI/LLM systems or modern observability stacks.
- Programming Languages: Strong proficiency in Python (for custom evals and telemetry scripts) and Go or TypeScript (for core platform services).
- Observability Ecosystem: Deep familiarity with OpenTelemetry (OTel), Prometheus, Grafana, and distributed tracing backends (Jaeger, Tempo, or Datadog).
- AI/LLM Observability Tools: Hands-on experience with tools like LangSmith, Langfuse, Arize Phoenix, Braintrust, or MLflow.
- AI Concepts: Solid understanding of LLM non-determinism, embedding/vector search mechanics, RAG pipelines, token economics, and concept/data drift.
Click on Apply to know more.