Now

I am currently living in New Delhi. This page serves as a record of what I am currently learning, building, and exploring in the world of engineering.

---

What I am Learning

I am focused on building a solid foundation across AWS cloud engineering, MLOps, and LLMOps.

• MLOps & Model Lifecycle: Data versioning with DVC, feature stores with Feast, tracking with MLflow, and scalable model serving with KServe & SageMaker AI.
• Generative AI & LLMOps: AWS Bedrock, Claude 3, pgvector for semantic retrieval, and function-calling guardrails.
• Docker and ECS: Multi-arch container builds, ECR registries, and serverless container deployments on AWS ECS Fargate.
• Terraform & IaC: Provisioning modular AWS infrastructure (VPC, ECS, ALB, IAM, RDS, CloudWatch).
• GitHub Actions CI/CD/CT: Automated build, test, deploy, and statistical drift-triggered retraining pipelines.
• Observability: Statistical drift metrics (Evidently AI, PSI), Prometheus/Grafana, and high-signal CloudWatch alerts.

---

What I am Building

• Indian Domestic Flight 'Buy vs. Wait' Advisory Platform: A full MLOps system with Feast feature store, MLflow registry, LightGBM, and sub-25ms FastAPI on ECS Fargate.
• Support & Sales Agentic RAG: Serverless GenAI assistant on AWS Bedrock (Claude 3, Titan) with pgvector HNSW indexing and SEBI compliance guardrails.
• Engineering Notes & Blogs: Writing practical, opinionated lessons from real production MLOps and cloud engineering setups.

---

What I am Improving

• Deployment Safety: Understanding rollback strategies, health checks, and zero-downtime deployments on ECS.
• Infrastructure Hygiene: Clean variables, modular Terraform, and remote state management.
• Observability Habits: Getting useful signal from CloudWatch logs instead of noise.

---

Current Priorities

1. Finish the ECS CI/CD project with full documentation and screenshots.
2. Complete the Terraform AWS Blueprint with dev and prod variable files.
3. Write blog posts that document what I learned from each project.

Last updated: