17 Nov
|
PRONATIVE AI
|
India
17 Nov
PRONATIVE AI
India
Job Title: ?? ????????????? ???????? (AgentGateway & ????????)
????? ??
At pronative.ai, we are building the Foundry for AI-Native Excellence. We operate a live, production-grade learning foundry powered by enterprise stacks (Anthropic Claude, Azure, AWS, and NVIDIA infrastructure). Our training delivery and multi-agent architectures run like high-availability software products, backed by mature platform engineering practices, automated agentic workflows, and robust AI governance.
???? ??????? & ?????? ???????????
This is a specialized, execution-focused ?? ??????? & ????????? ????—ideal for an engineer looking to own zero-code-change observability across LLM, Model Context Protocol (MCP), and Agent-to-Agent (A2A) traffic. Working directly alongside seasoned Platform Architects and Principal Engineers, you will configure, operate, and scale agentgateway.dev to capture distributed OpenTelemetry (OTel) traces and stream them directly into Langfuse. You will gain rare, practical exposure to ISO-certified (ISO 42001, 27701, 27017) AI management and cloud environments from day one.
???? ???????????????? (????????? ??? ??????? ????????????? ?????????)
- Pre-Training Gateway & Tracing Provisioning: Configure agentgateway.dev instances (Kubernetes Gateway API / standalone YAML) with native OpenTelemetry (OTLP) tracing policies to export spans directly to Langfuse (/api/public/otel/v1/traces) with zero application code changes. Manage Base64-encoded Langfuse authorization headers, TLS backend policies, sampling rules, and custom gen_ai.* semantic attributes for upcoming student cohorts and live AI pods.
- Live Delivery Traffic & Trace Monitoring: Monitor real-time LLM prompts/completions, MCP tool executions, A2A communications, and guardrail triggers flowing through AgentGateway into Langfuse dashboards. Track Time to First Token (TTFT), Time Per Output Token (TPOT), latency bottlenecks,
and routing failovers across providers using Langfuse analytics and AgentGateway access logs.
- Post-Training Audits, Token Cost Tracking & Governance: Audit per-user and per-session token consumption and model spend captured via AgentGateway's LLM traffic signals inside Langfuse. Execute post-cohort trace exports, retention cleanups, and OTel collector fan-out checks to maintain data privacy (ISO 27701) and AI compliance (ISO 42001).
??? ???? ????? ???'?? ???????
- AI Gateway & Routing: agentgateway.dev (Kubernetes Gateway API / AgentgatewayBackend / standalone YAML, MCP & A2A routing, CEL/prompt guardrails, built-in Observability UI).
- Observability & Telemetry: Langfuse (OTLP HTTP trace ingestion, Prompt & Cost Analytics, Session/User attribution), OpenTelemetry (OTel Collector fan-out, OTLP gRPC/HTTP, gen_ai semantic conventions).
- Cloud, Infrastructure & Automation: Kubernetes, Helm, Docker, Microsoft Azure / AWS, GitHub Actions, YAML/JSON configuration management.
????????? ???????
- Experience: 2 to 5 years of overall experience in Cloud Infrastructure, API Gateways, Platform Engineering, or Distributed Tracing, including a ??????? ?? ? ???? ?? ?????-?? ?????????? ?? ???????? (alongside practical exposure to AgentGateway).
- Operational Mindset: Robust attention to detail with a disciplined approach to maintaining gateway routing policies, OpenTelemetry collector pipelines, and observability SLAs.
- Core Technical Skills: Deep familiarity with agentgateway.dev tracing policies, OpenTelemetry (OTLP/gRPC/HTTP), Kubernetes Gateway API, and Langfuse trace/cost analytics.
- Eagerness to Innovate: A proactive, systems-thinking mindset—excited to work closely with AI architects to standardize gateway-level observability across enterprise agentic workloads.
Ready to lead AgentGateway & Langfuse Observability at the Frontier?
?????: ?????????.??
Apply directly or send your resume/Git profile to ???????@?????????.??
📌 AI Observability Engineer (India)
🏢 PRONATIVE AI
📍 India