AI and Document Search
LLM workflows, PDF processing, retrieval systems, citations, and practical document automation.
AI · Agents · Cloud · LLM
RAG pipelines, vector databases, LLM models, AI/ML agents devops.
AI tools to make sense of millions of documents. RAG pipelines based on millions of embeddings stored in vector databases. Run latest LLMs in efficient and cost effective mode.
Vector DBs and RAG pipelines
Solid aws cloud pipelines to ingest data.
Focused on AI applications, data engineering, cloud infrastructure, backend systems, and minimal user interfaces. More than a decade of experience in Python, AWS cloud.
Turn millions of messy documents, raw data, and manual workflows into searchable systems, useful dashboards, and automated cloud-based applications. Makes sense of huge corpus of data using RAG pipelines powered AI based decisions.
Group data and find unique characteristics like persona extraction etc using complex clustering algorithms.
Python
FastAPI
PostgreSQL
AWS
Azure
AI / ML
Milvus
PgVector
AWS Gpu
AI Agents
Kubernetes
Cloudwatch monitoring metrics
Prometheus
Grafana
OpenAI API
Services
AI/ML related programming and infrastructure services.
Data and Dashboards
SQL analysis, spend reports, operational metrics, dashboards, ETL scripts, and decision-ready summaries.
Cloud and Backend Systems
Github action, APIs, background jobs, AWS deployment pipelines, cloud infrastructure, containers, and secure automation.
Selected Projects
Four representative projects showing practical work in AI, analytics, cloud automation, and devops.
01
AI RAG Research Assistant
A document search and summarization workflow for extracting reliable answers from large PDF collections with page-level extract citations and concise final reports. Uses Milvus vector database holding large number of embeddings.
02
Persona extraction - AWS data pipeline
A corpus of more than millions of records was imported from freely available opensource census/sales data etc. A solid monitored pipeline was created using aws resources. Finally clustering algorithm was used to group similar characteristics and then RAG.
03
Cloud Automation Stack - Devops
Github action, Infrastructure automation for preview environments, background jobs, container workloads, secure deployment pipelines, and cloud cost control. AWS secrets manager, code checking, logging.
04
Orchestration - LLM locally
AWS step function used to orchestrate complex workflow and pipeline. CUDA drivers installed in AWS EC2 gpu for local execution of models. Monitoring using cloudwatch metrics using cloudwatch log parameters.
Make Contact
Available for short term and long term projects, consulting, technical reviews, AI prototypes, data dashboards, backend systems, and cloud devops automation work.