---
title: "Pranjal Verma (@pvcodes) | Hashnode"
description: "Data Engineer. Read the latest articles by Pranjal Verma on Hashnode."
url: "https://hashnode.com/@pvcodes"
---

@pvcodes

# Pranjal Verma

@pvcodes · India · Joined July 2020

Data Engineer

[Website](https://pvcodes.in/)

### [Local AI Stack for Productive Small Language Models](https://blog.pvcodes.in/local-ai-stack-for-productive-small-language-models?utm_source=hashnode&utm_medium=feed)

Making an SLM chat is easy; making it improve your day-to-day work is a tooling problem — here's the four-layer playbook.

7 min read · [blog.pvcodes.in](https://blog.pvcodes.in/local-ai-stack-for-productive-small-language-models?utm_source=hashnode&utm_medium=feed)

[Aug 29](https://hashnode.com/posts/local-ai-stack-for-productive-small-language-models/6a92b2a48973d388d8edb707)

### [Apache Iceberg Won the Format War - So Why Are We Rebuilding the Moat?](https://blog.pvcodes.in/apache-iceberg-won-the-format-war-so-why-are-we-rebuilding-the-moat?utm_source=hashnode&utm_medium=feed)

The format battle is settled. The new lock-in lives in catalogs, credentials, and policies your table format can't see.

7 min read · [blog.pvcodes.in](https://blog.pvcodes.in/apache-iceberg-won-the-format-war-so-why-are-we-rebuilding-the-moat?utm_source=hashnode&utm_medium=feed)

[Aug 23](https://hashnode.com/posts/apache-iceberg-won-the-format-war-so-why-are-we-rebuilding-the-moat/6a8a6e43e0a10a8289da6331)

### [Apache Iceberg: The Backbone of the Modern Open Lakehouse](https://blog.pvcodes.in/apache-iceberg-the-backbone-of-the-modern-open-lakehouse?utm_source=hashnode&utm_medium=feed)

How snapshots, REST catalogs, and a disciplined maintenance pipeline turn raw object storage into an operable multi-engine lakehouse

6 min read · [blog.pvcodes.in](https://blog.pvcodes.in/apache-iceberg-the-backbone-of-the-modern-open-lakehouse?utm_source=hashnode&utm_medium=feed)

[Aug 23](https://hashnode.com/posts/apache-iceberg-the-backbone-of-the-modern-open-lakehouse/6a8a6ae172d37ec77f8066ae)

### [Spark Performance Tuning Interview Prep Guide](https://blog.pvcodes.in/spark-performance-tuning-interview-prep-guide?utm_source=hashnode&utm_medium=feed)

Master Spark tuning for interviews: shuffle, partitioning, caching, joins, AQE, memory, Spark UI, and production scenarios for data engineering roles.

13 min read · [blog.pvcodes.in](https://blog.pvcodes.in/spark-performance-tuning-interview-prep-guide?utm_source=hashnode&utm_medium=feed)

[Aug 12](https://hashnode.com/posts/spark-performance-tuning-interview-prep-guide/6a7c3a02c4b28467692fba66)

### [Unstructured data becomes a first-class citizen: Databricks FILE type and the multimodal lakehouse](https://blog.pvcodes.in/unstructured-data-becomes-a-first-class-citizen-databricks-file-type-and-the-multimodal-lakehouse?utm_source=hashnode&utm_medium=feed)

Databricks FILE type turns documents, images, and video into governed, AI-ready lakehouse columns. How EXTERNAL vs MANAGED work, and why it matters.

8 min read · [blog.pvcodes.in](https://blog.pvcodes.in/unstructured-data-becomes-a-first-class-citizen-databricks-file-type-and-the-multimodal-lakehouse?utm_source=hashnode&utm_medium=feed)

[Aug 9](https://hashnode.com/posts/unstructured-data-becomes-a-first-class-citizen-databricks-file-type-and-the-multimodal-lakehouse/6a788a42e2c73b7894019295)