Contributed to the implementation of a multi-intent LLM conversational app.

Designed and implemented an LLM evaluation platform from the ground up, using a React frontend, FastAPI backend, and Snowflake for data pipelines.

Maintained and improved machine learning pipelines using Databricks MLflow, PySpark, and Airflow — covering the full lifecycle from feature extraction and model training to evaluation, prediction, and deployment.

Built, updated, and maintained AWS-based data pipelines leveraging Databricks, Spark, Scala, Airflow, Kafka, and Kubernetes.

Re-architected the existing Big Data architecture - Data pipelines, Data Lake, ML/Analytics, Reporting & Dashboard using AWS Databricks, AWS S3, AWS redshift, Mode and Airflow.

Designed and provisioned Databricks Lakehouse infrastructure using Terraform.

Designed and established the security architecture using CloudTrail, GuardDuty with email alerts, WAF with SQL injection and XSS attack rules, IAM role’s least privilege principle, AWS console and programmatic access MFA enforcement, and data validation and sanitization at the layers of frontend and backend.

Built, updated and maintained the Flask-based REST-ful APIs backend implementation with best practices for iOS/Android apps and Shopify website.

Designed and established the big data and machine learning cloud platform using Azure Databricks, Azure Data Lake Storage, and Azure Data Factory.

Designed and established the DevOps architecture using Azure DevOps pipeline, Azure DevOps agent, and Azure KeyVault.

Designed and established the application deployment architecture using Azure Kubernetes and Helm Chart.

Designed and established the performance testing automation tool (load testing, stress testing and endurance testing) with the integration of Azure DevOps with the JMeter.

Designed and developed an AI system to automate data collection process by extracting specific text information from the images using the computer vision's optical character recognition (OCR) methodology. Serverless-based AI deployment with the use of AWS Lambda function.

Designed and developed an AI system to automate the costly and time-consuming work order review process by using machine learning, deep learning, and natural language processing methodologies with the Koyos third party collaboration. Kubernetes-based AI deployment with the use of Istio/Amazon EKS.

Automated the set-up of a secure, multi-account AWS environment (AWS Organization) for running test/staging/production stages of software products.

Implemented a scalable, low latency, highly available, cost efficiency, secure and distributed SaaS cloud infrastructure with the use of the Serverless application framework, AWS Cloudformation, bash and python scripts.

Built a scalable ETL pipeline and a fast search engine with the use of AWS PostgreSQL RDS, AWS Glue, AWS Athena, Amazon S3, Amazon CloudWatch Event, AWS Lambda function and Amazon Elasticsearch.

Developed a data analytics and visualization platform that integrates the Looker BI tool with AWS PostgreSQL RDS, Amazon S3, AWS Athena, and custom visualizations (D3 and HighChart).

Secured SaaS cloud applications with AWS SecurityHub, CloudTrail, Macie, Inspector, GuardDuty, Config, Secret Manager, WAF, Shield Standard, VPC, KMS, IAM, MFA, S3 Bucket Policy, and Security Group.