Job Description 1: AI Architect

Experience: 9-12 years total in software engineering, platform engineering, data engineering, or enterprise/AI architecture, including at least 35 years of hands-on experience architecting or delivering production AI, ML or GenAI solutions. Must have led at least 2–3 full end-to-end AI/ML project lifecycles, from requirements and architecture through production deployment and support Reports to: DATA Platform Program Lead

Type: (Contact to hire)

Work mode - 5 day work from office - Gachibowli (Hyderabad)

Role summary

The AI Architect owns the end-to-end technical design of the OneEnterprise AI platform. You will define how natural-language questions flow through the semantic layer, ontology, knowledge graph and LLM/RAG services to produce accurate, explainable, and governed answers about cost, utilization and compliance. Will lead the AI engineering team and act as the technical bridge to cloud, finance, security, data and DevOps teams.

Key responsibilities

Architecture and design

- Define reference architectures for Retrieval-Augmented Generation (RAG), vector databases, LLM orchestration and agentic frameworks (e.g., LangGraph, CrewAI, AutoGen)
- Define how NLP, semantic layer, ontology, knowledge graph, LLM and RAG components integrate with the UI and data platform
- Design the OneEnterprise ontology: applications, services, infrastructure resources, accounts/subscriptions, cost centres, owners, SLAs, policies and compliance controls
- Choose and justify the technology stack (LLM providers or self-hosted models, vector store, graph database, orchestration frameworks, semantic layer tooling)
- Design hybrid retrieval combining knowledge-graph queries, semantic-layer metrics (text-to-SQL / text-to-metric) and vector search





Platform strategy

- Design shared internal AI platforms, APIs and developer experiences (DevEx) so product and data teams can build safely and efficiently
- Set reusable standards for prompt design, agent/tool orchestration and caching

MLOps and GenAIOps

- Oversee model lifecycle management, model registries, automated evaluations, latency tracking, and cost/token optimization in production
- Define the evaluation framework (accuracy, groundedness, latency, cost per query) and quality gates for release

Data and integration

- Architect ingestion from cloud billing and usage (AWS CUR, GCP Billing), OneEnterprise, monitoring/observability tools, tagging and policy engines
- Align the semantic layer with FinOps Framework concepts and FOCUS (FinOps Open Cost and Usage Specification) so metrics are consistent across providers
- Define data quality, lineage and entity-resolution approach for the knowledge graph

Responsible AI and governance

- Enforce data privacy, compliance, security-by-design, source attribution, and guardrails against system hallucinations
- Apply role-based access to cost and compliance data, and handle PII appropriately
- Ensure the solution meets enterprise security, audit and data-residency requirements

Leadership and cross-functional alignment

- Translate business goals into technical roadmaps, mentor engineers,



and guide executive stakeholders across multi-team programs
- Lead design and code reviews for the AI engineering team
- Partner with Cloud/Infra, Finance, Security, Data Engineering, UI and DevOps teams; translate business questions into technical requirements
- Own the AI roadmap, technical risks and architecture decision records

Required skills

- GenAI / LLMs: AWS Bedrock, Google Vertex AI, or open-source foundation models
- Data and RAG: vector databases (e.g., pgvector, Pinecone, Weaviate, Milvus), chunking, embeddings, hybrid search and re-ranking pipelines
- Frameworks: LangChain, LlamaIndex, Semantic Kernel, AutoGen, CrewAI or LangGraph
- Cloud and infrastructure: Kubernetes, Docker and cloud-native distributed architectures on GCP/AWS
- Knowledge graphs and ontology: Neo4j, Amazon Neptune, Stardog, GraphDB or similar; RDF/OWL, SPARQL and/or Cypher
- Semantic layer: dbt Semantic Layer, Cube, ; strong SQL and data modelling
- MLOps / GenAIOps: CI/CD, model registries, model and prompt versioning, evaluation and observability tooling
- Programming: robust Python; working knowledge of Java, Scala, TypeScript or Go
- Architecture: APIs and microservices; event-driven and streaming patterns

Education

Bachelor's or Master's degree in Computer Science, Engineering, Data Science or a related field

What success looks like in the first 12 months

1. Approved target architecture and ontology v1 within 30 days

2. MVP natural-language FinOps assistant live for pilot users within 3 months

3. Evaluation framework in place with agreed accuracy and groundedness targets

4. Production rollout across priority application portfolios with documented governance

📌 Artificial Intelligence Architect (Hyderabad)
🏢 Riktam Technology Consulting
📍 Hyderabad