I architect and build data platforms and AI products businesses depend on:

  • data pipelines processing over 5B+ records daily
  • analytics & AI dashboards that reduce reporting latency by 60%
  • operations automated end to end with AI agents, cutting execution time by half

real-time / batch pipelines - analytics - streaming - applied AI - data migration - warehousing - agentic workflows

Hero Image
Confluent Kafka local service stack

Proving Kafka Pipelines with the Confluent CLI: Publish and Subscribe

Description: Validate Kafka pipelines with console producer and consumer tools before writing application code.

  • kafka
  • confluent
  • streaming
  • cli
View Post
MLflow and Weights and Biases pipeline tracking

Weekly Rental Price Retraining: MLflow, W&B, and Hydra Pipeline

Description: Automated weekly rental price retraining with MLflow orchestration, W&B artifact tracking, and Hydra configs.

  • mlflow
  • wandb
  • mlops
  • python
View Post
AWS EMR Spark cluster architecture

Big data jobs: Spark ETL Design on AWS EMR

Description: Spark ETL on EMR in practice — S3 inputs, star-schema outputs, and deploy paths via console, CLI, and Boto3.

  • spark
  • aws
  • emr
  • data-engineering
View Post

Latest work and insights

Smolagents framework: Building Multi-Agents, Tools, and Memory for Order Processing
blog

2026-07-05

Smolagents framework: Building Multi-Agents, Tools, and Memory for Order Processing

smolagents multi-agent walkthrough — ToolCallingAgent specialists, tool trust boundaries, structured JSON handoffs, and sql db storage on a paper supply order pipeline

View
agentstoolssmolagentsmemory
agents
project

2026-06-27

Multi-agent system for customer service and sales operations

Orchestrating multiple agents to complete tasks involving handling customer inquiries, checking inventory status, providing accurate quotations, and completing transactions seamlessly

View
ai-orchestrationLLMsmolagentstool-calling
Snowflake data warehouse architecture
project

2026-05-24

A Snowflake Warehouse for Customer Reviews and Climate Analytics

Modeled Yelp reviews and GHCN-D climate data in a layered Snowflake warehouse from staging through ODS to star schema.

View
snowflakedata-warehousesqlanalytics
Club football micro-event data lake star schema
blog

2026-05-21

Micro-Event Data Lakes for Sports Analytics: Club Football on AWS

Sports analytics at scale — Spark, Airflow, and Redshift shaping 1M+ football micro-events from S3 into a star schema.

View
sparkairflowredshiftdata-engineeringsports

Specialization

  • Data Platforms (Kafka, Spark, Streaming)

  • Cloud Architecture (AWS, GCP, scalable systems)

  • AI Systems (RAG, Agents, LLM workflows)

  • Performance Engineering (low latency systems)

Impact

5B+ rows processed daily

50% faster data processing

5 days → 13 min ML retraining

98% NLP insight accuracy

70% better data accessibility

85% fewer pipeline breakages

Tech Stack

Python
Python
Java
Java
SQL
SQL
C++
C++
Kafka
Kafka
Spark
Spark
AWS
AWS
GCP
GCP
dbt
dbt
OpenAI
OpenAI
VectorDB
VectorDB
Airflow
Airflow
Terraform
Terraform

Let's Build Something Scalable