Skip to content
View ro-kannan's full-sized avatar

Block or report ro-kannan

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ro-kannan/README.md

Hi, I am Rohith

Analytics Engineering · BI & Visualization · AI Agents · Cloud Infrastructure

Tech Stack

Data Engineering & Warehouses: dbt · Apache Airflow · PySpark · AWS Glue · PostgreSQL · MS SQL · Snowflake · BigQuery · Redshift · Python AI/ML: RAG · LangGraph · OpenAI API · Ollama · pgvector · ChromaDB Cloud: AWS · Azure BI & Visualization: Power BI · Tableau · Looker Studio

Background

MBA in Big Data Analytics. 4 years at a data consulting firm delivering end-to-end data projects across BFSI, pharma, retail, and SaaS - spanning data engineering, BI, and analytics across US, UK, and Sri Lanka markets.

Managed engagements end-to-end - from scoping and SOW through delivery and closure - balancing hands-on technical delivery with client stakeholder management.

Work samples

Recurring patterns from client engagements, rebuilt as sample solutions with synthetic data.

Project Description Stack Demo
Multi-Agent Pipeline with Grounding Guardrails Four agents research, filter, write, and review real competitor trial data - when a citation can't be verified, the pipeline pauses for a human instead of publishing anyway. LangGraph · Groq · ClinicalTrials.gov API · Langfuse · Python -
RAG Hybrid Contract Search Naive vector search misses precise contract clauses - hybrid search (vector + BM25) closes the gap, measured head-to-head, with counterparty identity redacted before a cloud LLM ever sees it. pgvector · BM25 · Groq · Slack · Python -
Compliance RAG Baseline + Evaluation Fully local RAG over compliance policy documents - a chunking choice that quietly cuts a word in half turned out to drop a compliance-critical clause. Ollama · ChromaDB · RAGAS · DeepEval · Python -
Lending Data MCP Server A governed MCP layer wrapping a lending data API - the same access rules apply whether a human, a dashboard, or an AI agent is asking. MCP · FastAPI · PostgreSQL · Python -
dbt Metrics Layer - Pharma KPI Governance Five teams, one KPI, five different answers - a Semantic Layer that enforces a single definition. dbt · MetricFlow · DuckDB · Streamlit Live
Event-Driven Order Ingestion Pipeline File upload triggers a serverless pipeline - validates two order sources, quarantines bad batches, loads idempotently. AWS EventBridge · Step Functions · Lambda · RDS PostgreSQL · Python -
Airflow Retail Order Pipeline - AWS S3 Dead-Letter Omnichannel orders from two sources hit a quality gate - clean rows load to analytics, bad rows are quarantined to S3 before they reach the database. Apache Airflow · AWS S3 · PostgreSQL · Docker · Python -
Multi-Consumer AI Data Gateway The controlled layer between AI agents and your database - per-consumer auth, rate limits, and validated SQL execution. FastAPI · SQLAlchemy · PostgreSQL · Python -
SCD Type 2 - Lending Dimension Historical reports silently return wrong numbers when a dimension attribute changes - because the join returns today's value, not what it was at the time. SCD Type 2 fixes this. dbt · PostgreSQL · Python -
On-Prem to Cloud Migration Full on-prem to cloud migration: incremental extraction, automated on a schedule, and a dbt transformation layer handling the migration seamlessly. AWS · Snowflake · EC2 · dbt -
Analytics Engineering - E-commerce Pipeline Transforms raw e-commerce transaction logs into reliable dbt marts - incremental loads, 16 data quality tests, and automatic alerts when source data goes stale. dbt · PostgreSQL · Python -

Pinned Loading

  1. multi-consumer-ai-data-gateway multi-consumer-ai-data-gateway Public

    FastAPI data gateway — per-consumer API key auth, per-key rate limiting, and validated SQL execution for AI agents. Built over a PostgreSQL lending schema.

    Python

  2. aws-snowflake-onprem-migration aws-snowflake-onprem-migration Public

    On-prem to cloud migration: PostgreSQL → S3 lifecycle policy → IAM least-privilege → EC2 → Snowflake external stage + COPY INTO → dbt. Zero SQL dialect changes.

    Python

  3. dbt-metricflow-pharma-analytics-governance dbt-metricflow-pharma-analytics-governance Public

    5 brand teams, 5 different net revenue numbers — same raw data. Governed dbt metrics layer eliminates divergence across 5 pharma commercial scenarios.

    Python

  4. airflow-retail-pipeline-aws airflow-retail-pipeline-aws Public

    Two-source Airflow pipeline: store ERP CSV + e-commerce API → quality gate → PostgreSQL analytics or S3 dead-letter quarantine

    Python

  5. lending-data-mcp-server lending-data-mcp-server Public

    Python

  6. rag-hybrid-contract-search rag-hybrid-contract-search Public

    Python