AI and Document Search
LLM workflows, PDF processing, retrieval systems, citations and practical document automation.
AI · Agents · Cloud · LLM
Make vast data useful.
RAG pipelines, vector databases, LLM applications and AWS infrastructure engineered to search millions of records, automate complex workflows and run reliably in production.
# production AI/data workflow
source = "millions_of_records"
embeddings = create_vectors(source)
index = milvus.upsert(embeddings)
workflow = step_functions.run({
"api": "AWS Lambda",
"retrieval": "RAG",
"observability": "enabled"
})
# secure · monitored · cost-aware
▌
100M+Vector records administered
60+AWS accounts supported
Python10+ years Hands-on senior engineering
End-to-endArchitecture through operations
Vector Stores & RAG pipelines
Systems that turn complexity into decisions.
Senior AWS Architect, hands-on AWS DevOps engineer and Senior Python Programmer focused on serverless APIs, AI applications, data engineering and production cloud systems. Recent work adds more than a year of GenAI, RAG and LLM delivery to a long track record in Python and AWS.
Millions of messy documents, raw records and manual workflows become searchable systems, useful dashboards and automated cloud applications. Clustering, retrieval and AI-assisted analysis reveal patterns, personas and actionable signals.
AI & RetrievalRAG · Agents · OpenAI - Claude
Vector DataMilvus · pgvector
CloudAWS · Azure · Kubernetes - Terraform
EngineeringPython · Lambda Powertools · CDK TypeScript
ObservabilityCloudWatch · Grafana - Prometheus
Services
Practical engineering, Production.
AI/ML programming and infrastructure services built around measurable outcomes, maintainable systems and production operations. Ai guardrails
Large-Scale Data & ETL
AWS Glue, PySpark, S3/Lambda ingestion, Fargate, AWS Batch, GPU processing, complex SQL and production data pipelines.
AWS Architecture and Serverless
Senior AWS architecture, Terraform, Python Lambda APIs, AWS Lambda Powertools, CDK TypeScript, DevOps pipelines and cost-aware production operations.
Selected projects
Built for real prod workloads.
Representative work across AI, analytics, serverless APIs, cloud automation and production DevOps.
PROJECT 01 · FEATURED
Open WebUI Customization & AI Workflows
Source-level Open WebUI customization with Python/FastAPI, Svelte integration, product and persona workflows, review-backed RAG analytics, AWS orchestration and fork upgrades.
Explore Open WebUI customization
PROJECT 02 · FEATURED
pgvector, Milvus & 100M-Record Vector Databases
Large-scale vector data engineering across AWS RDS PostgreSQL, Milvus on EKS, Helm customization, OpenAI embeddings, AWS Batch and production RAG search.
Explore vector database engineering
PROJECT 03 · FEATURED
AWS Glue, PySpark & Large-Scale ETL
AWS-native data processing for large datasets using Glue, PySpark, S3-triggered Lambda, ECS Fargate, AWS Batch, EC2 GPU, Python and complex SQL.
Explore the data-engineering portfolio
PROJECT 04 · FEATURED
Shopify Custom Programming & AI
Simple theme changes through advanced custom apps, Shopify Functions, extensions, API integrations, AWS backends and practical AI commerce automation.
Explore Shopify programming services
PROJECT 05 · FEATURED
AWS Lambda, API Gateway & Python serverless Systems
Senior AWS architecture and hands-on Python delivery for secure serverless APIs, AWS Lambda Powertools, CDK TypeScript, authorizers, Step Functions and recent GenAI/RAG integrations.
Explore the full portfolio
PROJECT 06
LLM Orchestration on AWS
AWS Step Functions coordinating GPU instances, local model execution and background processing, with CUDA automation and CloudWatch-based monitoring.
Make contact
Need a senior engineer who can own the full system?
Available for short- and long-term projects, consulting, technical reviews, AI prototypes, data dashboards, backend systems and cloud DevOps automation.
I’m open to very short-term assignments as well as long-term programming associations. Engagements can be arranged through fixed milestone payments or weekly payment schedules, and I’m also open to a small demo or short test job when needed.