Understanding of technology and services related to the domain.
Key Responsibilities
End‑to‑End Solution Design & Development
- Translate business and operational requirements into architecture, user stories, and technical designs for Dark NOC features.
- Build full‑stack AI capabilities : model integration (LLMs/NLP), orchestration logic, APIs/services, and workflow automation.
- Implement RAG pipelines (data ingestion, chunking, embeddings, vector search) and tool/function calling for autonomous actions.
Model Integration & Prompt Engineering
- Integrate LLMs/NLU/ASR/TTS providers with robust adapters, retries, timeouts, and fallbacks.
- Design and maintain prompts, system policies, and tool schemas ; evaluate and refine prompts for accuracy and reliability.
- Implement guardrails (policy enforcement, PII masking, safety filters) and quality evaluation (e.g., RAG ground truth checks).
Data Engineering for Dark NOC
- Build data ingestion & transformation for logs, alerts, tickets, and knowledge bases.
- Maintain feature/knowledge freshness SLAs and data contracts with upstream systems.
- Integration with New Relic, Service Now, Email, chat, REST API for end-to-end automation.
Testing, Quality & Evaluations
- Implement unit/integration/e2e tests , plus AI evaluations (groundedness, hallucination, toxicity).
- Create offline and shadow/A‑B evaluations for prompts, models, and RAG changes before production rollout.
- Define acceptance criteria with the Project Manager; maintain a robust regression suite.
CI/CD & Operations‑Ready Builds
- Set up CI/CD pipelines with canary/blue‑green releases, automated rollbacks, and migration/versioning for prompts, models, and indexes.
- Containerize services (Docker) and deploy to Kubernetes with observability hooks and resource limits.
- Produce runbooks and operational toggles (feature flags, kill‑switches, fallback modes).
Collaboration & Delivery Management
- Work closely with the Project Manager on scope, estimations, milestones, and risk tracking.
- Partner with platform, infra, and data teams to unblock dependencies and align environments and SLAs.
- Provide clear documentation (designs, APIs, runbooks, evaluation results) and demo increments to stakeholders.
Production Support (L3‑Level)
- Support pre‑prod validations and production rollouts; analyze incidents with traces/logs and drive code fixes.
- Own RCA for code/config issues and convert findings into tests, guardrails, and automation.
KPIs
- Feature Delivery Predictability: % of committed Dark NOC stories delivered per sprint / quarter.
- Deployment Frequency: Target : 2–4 per week
- Change Failure Rate: Target: 8–10%
- Defect Escape Rate: Target: 10–15%
AI Effectiveness: Target:- 90%
- Auto‑Remediation Success: Target: 60% for repeatable issues
Preferred Trainings/Certifications
- Certification - Generative AI with Large Language Models or LLMOps
- Certified in C# (.NET) or Python
- Certification or deep knowledge of SDLC, Agile, or DevOps methodologies
