Lalit Moharana - Builder of systems that think
Independent engineer and CTO at StriveSteam. I build agentic AI, LLM infrastructure, and data-native platforms for messy, real-world data: emails, PDFs, SAP exports, social streams. Creator of OmniQuery, the Agent Platform, and an AI research newsletter. AWS Community Builder in Data Engineering (2025, 2026).
Recent writing
- Querying the lake, without leaving the warehouse — — A warts-included build of Amazon Aurora PostgreSQL's direct query support for Apache Iceberg - real S3 lake, real Serverless v2 cluster, every command and every error message, ending in one query that spans five years of lake history and this morning's live rows.
- Complete Guide to All 23 Design Patterns in Agentic Python Systems — — A field guide to Gang-of-Four design patterns applied to AI agent backends - creational, structural, and behavioral - with Python examples pulled straight from production code.
- ZooKeeper's farewell: Apache Kafka enters a new era with KRaft — — Why Kafka is finally shedding ZooKeeper, what KRaft mode changes operationally, and how to plan a migration.
- DuckDB Unleashed: distributed data processing with SmallPond — — How SmallPond turns embedded DuckDB into a distributed compute engine - and where that pattern is going next.
- The rise of single-node processing engines (and their distributed cousins) — — DuckDB, Polars, and Datafusion made single-node analytics fast again. Here is what that shift means for data platform design.
- Harnessing Amazon S3 Tables with PySpark locally — — Running PySpark against Amazon S3 Tables from your laptop - configuration, catalog setup, and gotchas.
- Building real-time data lakes with Apache Flink and Apache Paimon — — A hands-on guide to streaming CDC into an Apache Paimon lakehouse with Flink SQL - including changelog semantics, compaction, and query patterns.
- RAG Fusion: overcoming RAG limitations with the wisdom of King Bhoj's court — — Why single-query RAG breaks down at retrieval time, and how RAG Fusion multiplexes queries to catch what single-shot misses.
- Simplifying serverless: multiple endpoints in one Python Lambda with APIGatewayRestResolver — — How AWS Lambda Powertools' APIGatewayRestResolver turns one Python Lambda into a full mini-API - routes, decorators, validation.
- Mastering CDC with Hudi: managing trains at an Indian railway station — — A field metaphor - Indian railway stations - for wrapping your head around Change Data Capture on Apache Hudi.
- AWS Lambda idempotency for Python devs, using Powertools — — Why every serverless workflow eventually needs idempotency, and how AWS Lambda Powertools makes it a decorator away.
- Demystifying Spark cluster configuration: a desi data engineer's guide — — Sizing executors, cores, and memory for Apache Spark - explained through the lens of Indian street food.
- Introducing Project Nessie: revolutionizing data lake management — — Git-like versioning, branching, and merging for your data lake - what Project Nessie changes about how we think about lakehouse governance.
- Migrate data like a boss with Airbyte — — Setting up Airbyte with Docker Compose and building a pipeline to migrate between two Postgres databases.
- Vector similarity on AWS Neptune Analytics — — Combining property graphs with vector embeddings on Neptune Analytics to surface hidden connections.
- A beginner's guide to containerizing FastAPI + React with Docker — — A step-by-step walkthrough for wrapping a FastAPI backend and a React frontend into Docker containers and orchestrating them locally.
- Setting up a 3-node Apache NiFi cluster — — Standing up a production-shaped 3-node Apache NiFi cluster - ZooKeeper, TLS, load balancing, and site-to-site.
- Creating custom NiFi processors and Dockerizing them — — A step-by-step guide to writing your own Apache NiFi processor in Java and shipping it inside a Docker image.
- Knowledge graphs and AI: illuminating insights from PDFs — — Building a knowledge graph from unstructured PDFs and layering AI on top to surface answers.
- Deploying a Python model to AWS Lambda with Jenkins — — A minimal CI pipeline that packages a Python model, zips its dependencies, and deploys to AWS Lambda on every merge.