🎓 Ph.D. in IoT | ☁️ Certified GCP ML Engineer & AWS Solutions Architect
📍 Pune, India

I build scalable, automated, and secure cloud infrastructure and bridge the gap between Data Science and Operations by engineering enterprise-grade MLOps pipelines.

• Technical Lead & Cloud Architect with 12+ years of experience in distributed systems and cloud engineering.
• 🤖 Transitioned deeply into MLOps & LLMOps, specializing in operationalizing machine learning models using GCP Vertex AI and Kubeflow.
• ⚙️ Passionate about FinOps (reducing cloud spend by up to 30%), GitOps (ArgoCD), and Zero-Trust Security across multi-cloud environments.
• 🚀 Currently building scalable Monorepos featuring Agentic AI, RAG architectures, and real-time streaming ML pipelines for the financial sector.

A comprehensive Monorepo demonstrating production-grade machine learning architectures on GCP, transitioning from traditional MLOps to modern LLMOps (Gemini).

Architecture Flow Data Ingestion (Pub/Sub + Dataflow) $\rightarrow$ Feature Store $\rightarrow$ Orchestration (Vertex Pipelines/Airflow) $\rightarrow$ Serving (Vertex Endpoints) $\rightarrow$ Monitoring (Looker Studio)

Key Features

• AML Fraud Detection: Real-time behavioral scoring processing massive transaction volumes with <100ms latency using Vertex AI Endpoints.
• Automated Credit Analysis: LLMOps pipeline utilizing Document AI to extract financial P&L statements and Gemini 1.5 Flash for zero-shot credit risk memos.
• Algorithmic Pricing: Orchestrated ephemeral Dataproc (Spot VMs) clusters to process high-frequency market data, cutting compute costs by 80%.
• Loan Loss Provisioning (IFRS-9): Leveraged BigQuery ML for in-warehouse model training, eliminating data egress costs.

Tech Stack
GCP • Vertex AI • Kubeflow • Terraform • Gemini • BigQuery • Python

🔗 Repository: https://github.com/rathoddt/mlops-with-vertex-ai

An infrastructure-as-code and deployment automation project designed to orchestrate clusters globally while maintaining 99.99% uptime and enforcing strict cloud governance.

Pipeline Flow IaC (Terraform) $\rightarrow$ CI/CD (GitHub Actions/Jenkins) $\rightarrow$ GitOps (ArgoCD) $\rightarrow$ Deployment (GKE/EKS) $\rightarrow$ Observability (Prometheus/Grafana)

Key Features

• Built credential-less multi-cloud authentication via Workload Identity Federation (AWS $\leftrightarrow$ GCP).
• Achieved a 63% reduction in deployment time by automating 100+ CI/CD pipelines.
• Engineered a complete observability stack reducing MTTR for critical incidents by 23%.

Tech Stack
Terraform • Kubernetes (GKE/EKS) • Docker • ArgoCD • Prometheus/Grafana

Cloud Platforms
Google Cloud Platform (GCP) • Amazon Web Services (AWS) • Microsoft Azure

DevOps & IaC
Terraform • Kubernetes (GKE, EKS, AKS) • Docker • Helm • Istio

MLOps & AI
Vertex AI • Kubeflow • MLflow • TensorFlow Extended (TFX) • RAG • BigQuery ML • Gemini API

CI/CD & Automation
GitHub Actions • GitLab CI • Azure DevOps • ArgoCD (GitOps) • Jenkins

Observability & Scripting
Prometheus • Grafana • Loki • Python • Bash • PowerShell

⭐ If you find my cloud architectures or MLOps pipelines helpful, feel free to star the repositories!