AI Architect

Tata Consultancy Services Limited; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

Tata Consultancy Services Limited is seeking experienced professionals in Singapore to design and deploy large language model solutions and RAG pipelines. You will train and fine-tune models, implement vector databases, and build scalable LLM-enabled applications within enterprise architectures.

You will work with cloud AI services, Kubernetes, and MLOps practices to ensure secure, observable, and compliant deployments across business units.

Qualifications

• Hands-on Python and at least one of Java or .NET.
• Experience building LLM solutions and RAG pipelines.
• Ability to train or fine-tune custom LLMs, not just use APIs.
• Experience with vector databases and embeddings (FAISS, Pinecone, pgvector).
• Familiar with Delta Lake/Lakehouse and modern data platforms.
• Experience with multi-turn dialogue systems and state handling at scale.
• Exposure to cloud AI services (Azure, AWS, Google) and containerization (Docker/Kubernetes).
• Knowledge of MLOps (MLflow, Kubeflow, CI/CD, model registries).
• Understanding of enterprise integration, API-first design, microservices.
• Knowledge of security, governance, Responsible AI, data privacy, threat modeling.

Skills

Python

Java

.NET

LLM pipelines

Prompt engineering

LangChain

LlamaIndex

RAG

Evaluation metrics

Red-teaming

Fine-tuning

Vector databases

FAISS

Pinecone

pgvector

Delta Lake

Lakehouse

Kubernetes

Cloud AI services

Azure OpenAI

Azure AI Search

AWS Bedrock

SageMaker

Vertex AI

Model training

MLOps

CI/CD

Model registries

Observability

Security & governance

Responsible AI

Data privacy

Threat modeling

Tools

LangChain

LlamaIndex

FAISS

Pinecone

Delta Lake

Lakehouse

Kubernetes

Azure OpenAI

Azure AI Search

AWS Bedrock

SageMaker

Vertex AI

MLflow

Kubeflow

CI/CD tools

Job description

• Hands-on experience with Python preferred, plus Java or .NET; strong software engineering knowledge and familiarity with design patterns.
• Experience building LLM solutions, including RAG pipelines, prompt engineering, agent frameworks such as LangChain or LlamaIndex, and evaluation using gold sets, metrics, and red-teaming.
• Proven experience training or fine-tuning custom LLMs, not only consuming managed APIs.
• Expertise with vector databases and embeddings, such as FAISS, Pinecone, and pgvector; feature stores; and modern data platforms such as Delta Lake or Lakehouse.
• Hands-on experience with multi-turn dialogue systems, conversational state handling, and intent orchestration at scale.
• Exposure to cloud AI services, including Azure OpenAI, Azure AI Search, AWS Bedrock, SageMaker, and Google Vertex AI; experience with containerization and Kubernetes.
• Knowledge of LLMOps and MLOps practices, including MLflow or Kubeflow, CI/CD, model registries, deployment strategies, monitoring for drift and toxicity, and observability.
• Strong understanding of enterprise integration patterns, including SOA, ESB, EAI, ETL, API-first design, event streaming, and microservices.
• Knowledge of security and governance, including Responsible AI, data privacy (PII and PCI), model risk management, access controls, secrets management, and threat modeling.
• Experience documenting conceptual, logical, and physical architectures; capability and process inventories; and clear architectural blueprints.
• Preferred certifications include TOGAF, Azure/AWS/GCP Cloud Architect, or Responsible AI certifications; exposure to Agile, Scrum, and DevSecOps is beneficial.