Febrianto Arif

• • Designed and automated robust end-to-end MLOps pipelines on AWS and GCP using SageMaker, Vertex AI, and Terraform to deliver scalable, reliable model training, deployment, and resource management.

• • Data Filtering and cleaning of health information system application.

• • Containerized machine learning models with Docker and used Kubernetes for orchestrated, zero downtime rollouts with blue/green and canary deployment strategies.

• • Built and maintained real-time monitoring with Prometheus, Grafana, and CloudWatch, tracking model health and triggering alerts for performance or SLA issues.

• • Developed automated data and model drift detection pipelines using MLflow and Evidently AI, allowing for rapid retraining and real-time integration of model updates when data shifted.

• • Controlled batch training jobs and recurring workflows with Apache Airflow and Kubernetes, keeping full audit trails and transparency across the ML lifecycle.

• • Integrated feature stores with Feast to standardize and reuse features, improving experiment speed and data accuracy organization-wide.

• • Improved CI/CD with Jenkins and GitHub Actions, embedding robust testing, security scans, and compliance checks for fast and error-free ML deployments.

• • Led company-wide adoption, deployment, and fine-tuning of LLMs and GenAI solutions (GPT-4, Llama, Claude, Gemini) for nuanced tasks like semantic search, content generation, and knowledge management.

• • Built and optimized Retrieval-Augmented Generation (RAG) pipelines using LangChain, Hugging Face Transformers, and Pinecone, delivering secure, context-aware Q&A and retrieval systems.

• • Oversaw responsible adaptation and fine-tuning of LLMs with company data, enforcing strong governance, tracking, and compliance across all experiments.

• • Built and maintained robust REST and gRPC APIs for scalable, secure access to AI and GenAI services on both internal and customer-facing platforms.

• • Led the design and deployment of full-stack Python RAG pipelines using GPT-4, Qdrant, and LangChain, reducing support resolution time by 32%.

• • Designed multi-agent triage workflows using Agentic AI patterns (LangGraph + LangChain agents) for autonomous routing, retrieval, and clinical decision support.

• • Mentored junior ML engineers and new hires on GenAI tooling, best practices, and evaluation methodology.

• • Fine-tuned LLMs for triage and Q&A tasks using PEFT and LoRA.

• • Deployed asynchronous LLM inference routing using Ray Serve across multiple models with load balancing.

• • Used Ray Tune for distributed hyperparameter tuning on transformer-based models.

• • Built scalable data pipelines with Databricks and Delta Lake to prepare multi-source healthcare data.

• • Integrated evaluation harnesses with NeMo Guardrails and Rebuff for hallucination monitoring.

• • Integrated clinical document analysis using OpenCV and Tesseract for OCR, enabling vision-language triage pipelines in tandem with GPT-4-based summarization models.

• • Deployed scalable LLM and RAG pipelines on GCP Vertex AI Pipelines using custom containers with CI/CD, monitoring, and model versioning.

• • Managed secure data pipelines on Databricks + GCS + Delta Lake, and integrated with BigQuery for downstream analytics.

• • Experimented with Neo4j Healthcare

• • Developed distributed training pipelines using Databricks on AWS to process 100M+ record fraud datasets, enabling real-time risk scoring and dynamic rule evaluation.

• • Prototyped Python-based LLM-serving stack using Ray Serve + Triton Inference Server for GenAI demos.

• • Built SageMaker Pipelines for end-to-end model development lifecycles, covering feature engineering, training, evaluation, deployment, and A/B testing.

• • Integrated Neo4j with AWS Glue and SageMaker pipelines to perform graph-based feature generation and link analysis for anomaly detection use cases.

• • Integrated Ray Train into EC2 clusters for image classification workflows using CNNs and PyTorch, supporting product classification and medical image analysis PoCs.

• • Led PoCs for real-time document OCR and object detection with OpenCV, YOLOv5, and Detectron2, integrated with SageMaker endpoints.

• • Integrated AWS CloudWatch and Prometheus exporters to monitor pipeline performance, with Grafana dashboards for training jobs, endpoint latency, and error rates.

• • Built GitHub Actions + Terraform pipelines for CI/CD and model deployment.

• • Unified ML experimentation across teams via shared MLflow registry.

• • Consulted with AWS clients on feature store and vector DB adoption.

• • Benchmarked GenAI inference latency across container runtimes and backends.

• • Delivered live notebooks and reference repos to support AWS workshops