AI Architect – AWS ML Solution (Hands-On)

Role Details

- Experience: 10–12 years
- Primary Tech/Domain: AWS
- Must have strong platform engineering skills.

Overview & Expectations

Role Summary:

Design, build, lead, and deliver high-impact AI solutions. Own execution excellence with measurable business value, technical depth, and governance.

Key Outcomes (06–12 months):

- Ship production-grade solutions with clear ROI, reliability (SLOs), and security.
- Establish engineering standards, pipelines, and observability for repeatable delivery.
- Mentor talent; uplift team capability through reviews, playbooks, and hands-on guidance.
- Needs to build a successful AI platform that can be used to build AI applications across different domains

Responsibilities:

- Translate business problems into well-posed technical specifications and architectures.
- Lead design reviews, prototype quickly, and harden solutions for scale (1M+ users / high QPS).
- Build automated pipelines (CI/CD) and model/data governance across environments.
- Define & track KPIs: accuracy/latency/cost, adoption, and compliance readiness.
- Partner with Product, Security, Compliance, and Ops to land safe-by-default systems.
- Must have built an AI platform for AI applications across different domains.

Technical Skills: (Most of the levels mentioned below, but must have agentic AI skills)

- Platform: AWS SageMaker,; Kubernetes + Docker for portable inference
- MLOps: MLflow registry, Kubeflow Pipelines, CI/CD (GitHub Actions/Jenkins), canary/champion-challenger
- Monitoring: SageMaker Model Monitor (data/quality drift), monitoring, CloudWatch/Prometheus
- Serving: TensorFlow Serving, TorchServe,



FastAPI/GRPC, autoscaling, low-latency optimizations
- Feature stores & data: Feast, Databricks Feature Store; batch vs. streaming pipelines
- Should have a solid practical knowledge of Machine learning and statistics.
- Security: secret management, network isolation, encryption-at-rest/in-transit, compliance logging
- AWS Bedrock & LLM Integration – Deploy and customize foundation models (GPT, Claude, Titan) using Amazon Bedrock, including prompt engineering and fine-tuning strategies.
- RAG Architecture with AWS Services – Implement Retrieval-Augmented Generation using Amazon Kendra for semantic search and OpenSearch for vector embeddings.
- Agentic AI Orchestration – Design multi-agent workflows leveraging AWS Lambda, Step Functions, and integration with LangChain/CrewAI for event-driven orchestration.
- MLOps & Deployment on AWS – Build pipelines with SageMaker Pipelines, manage model registry, and enable CI/CD using CodePipeline and CodeBuild.
- Security & Compliance on AWS – Apply AWS IAM, KMS, and Secrets Manager for secure access; ensure compliance with GDPR, HIPAA, and SOC using AWS governance frameworks.
- Design agentic systems using AWS Lambda and Step Functions for orchestration and state management.




- Integrate multi-agent workflows with Amazon EventBridge for event-driven architectures.
- Secure agent operations using AWS IAM, KMS, and Secrets Manager for identity and encryption.
- Connect agents to Amazon Comprehend, Polly, and Lex for NLP and conversational capabilities.
- Implement observability with CloudWatch, X-Ray, and distributed tracing for agent workflows.

Architecture & Tooling Stack:

- Must have designed and built at least 3 Agentic AI solutions in Azure.
- Must have platform engineering experience.
- Source control & workflow: Git, branching standards, PR reviews, trunk-based delivery.
- Containers & orchestration: Docker, Kubernetes, Helm; secrets, configs, RBAC.
- Observability: logs, metrics, traces; dashboards with alerting & on-call runbooks.
- Data/Model registries: metadata, lineage, versioning; staged promotions.

Performance & Reliability:

- Define SLAs/SLOs for accuracy, tail latency , throughput, and availability.
- Capacity planning with autoscaling; load tests; cache design; graceful degradation.
- Cost controls: instance sizing, spot/reserved strategies, storage tiering.

Security & Compliance:

- IAM, network isolation, encryption (KMS), secret rotation.
- Threat modeling, dependency scanning, SBOM, supply-chain security.
- Domain-regulatory controls (PCI DSS, HIPAA) where applicable; audit readiness.

Qualifications:

- Bachelor’s/Master’s in CS/CE/EE/Data Science or equivalent practical experience.
- Solid applied programming in Python; familiarity with modern data/ML ecosystems.
- Proven track record of shipping and operating systems in production.

📌 AI Architect - AWS ML Solution (Bengaluru)
🏢 Bandhan Technologies
📍 Bengaluru