--- date: January 16, 2026 · slug: inside-git ---
In the previous post, we learnt [How Git manages version control](post.html?post=git-essentials) through its internal structures. In this continuation, we will delve deeper into h
cat full-post.md →1
vibhavari@portfolio:~$ neofetch
vibhavari@portfolio
─────────────────────────
RoleData Engineer · Backend Engineer
Focuspipelines & platforms · APIs & backend services
DataApache Spark · Kafka · Databricks · Terraform · SQL
BackendPython/TypeScript · FastAPI/Express · Node.js · AWS Cloud
LocationFrance / remote-friendly
Siteportfolio · writing · open notes
2
vibhavari@portfolio:~$
3
vibhavari@portfolio:~$
vibhavari@portfolio:~$ cat metrics.txt
pipeline runtime10h → 2hMews · Databricks
compute cost (DBUs)↓ 60–75%Mews · platform
profile update latency↓ 75–80%AgileLab · Flink/Kafka
query performance< 3s p99AgileLab · Spark/Presto
LLM chat backendFastAPILezzCo · backend
WhatsApp integrationTwilio webhooksLezzCo · API
daily data processed30GB+Deloitte · ETL
workflow efficiency↑ 40%Deloitte · orchestration
technical posts11~/blog
~/experience.git — git log --graph --career5 commits · main · scroll ↓
*
commit a3f9e21 (HEAD → main, tag: current) · tag: data
Data Engineer - Platform @ Mews, France
Date: July 2025 - Present
- Redesigned platform architecture to adopt configuration-driven, Infrastructure-as-Code patterns using Terraform, standardizing SQL Warehouse management and access controls to eliminate configuration drift and enable declarative, low-risk permission changes on Databricks.
- Optimized critical data pipelines, slashing runtime from 10+ hours to 2 hours while reducing compute costs (DBUs) by 60-75%.
- stack: Databricks, Python, SQL, Terraform
*
commit 7c41b08 · tag: backend
Backend Engineer @ LezzCo
Date: May 2025 - Present
- Designed and implemented backend for LLM-based chatbot that powers product conversations.
- Integrated Shopify context so customers can ask about products and get relevant responses.
- Enabled WhatsApp-based customer chat via Twilio webhooks with secure request validation.
- Built the API proxy and backend flows for authentication, chat routing, and metrics tracking.
- stack: AWS Cloud, Python, FastAPI, AWS Bedrock
*
commit e98d3aa · tag: data
Data Engineer @ CyberSecurity Client
Date: May 2025 - July 2025
- Architected cloud-native data pipeline integrating four disparate systems (Runn.io, ConnectWise, QuickBooks, Hubspot) using Cloud Functions and BigQuery with near real-time synchronization.
- Designed unified data models powering Looker dashboards that accelerated decision-making across business functions.
- stack: Google Cloud Functions, BigQuery, Python, API Integration
*
commit 1b07f44 · tag: data
Data Engineer @ AgileLab, France
Date: 2021 - 2024
- Spearheaded cross-functional collaboration with data engineers and scientists to deploy advanced NLP models on AWS cloud, cutting deployment time by 50%.
- Developed real-time data pipelines using FlinkSQL and Kafka CDC to process hundreds of GBs of customer data, reducing profile update times by 75-80%.
- Ensured GDPR compliance through precise anonymization rules with Livy, Spark, and Presto, achieving sub-3-second query performance while maintaining enterprise-grade data security.
- stack: Apache Spark, FlinkSQL, Kafka, Presto, AWS
*
commit 0000001 · tag: data
Big Data Engineer @ Deloitte, Bangalore
Date: 2018 - 2021
- Building and optimizing large-scale ETL infrastructure using Hadoop, Spark, and Hive.
- Created automated orchestration frameworks and comprehensive logging systems, achieving 40% efficiency gains in workflows and issue resolution while processing 30GB+ of data daily to drive business intelligence.
- stack: Hadoop, Apache Spark, Hive, ETL, Python, Scala, Workflow Orchestration
~/blog — find . -name '*.md' | head -5filetype: markdown
--- date: January 13, 2026 · slug: git-essentials ---
If you have ever played a video game, you know the anxiety of facing a difficult boss. What is the first thing you do before walking through that boss door? You save your game. W
cat full-post.md →--- date: January 10, 2026 · slug: why-version-control-exists ---
## Why Version Control Exists - The Pendrive Problem **Analogy: The Pendrive Problem** Let's imagine a world where version control systems like Git do not exist. Instead, develo
cat full-post.md →--- date: Mar 21, 2025 · slug: automate-unit-tests-using-ci-cd ---
In this blog, we’ll explore how to set up a complete CI/CD pipeline using `Jenkins`, `pytest`, and `Terraform` to automate unit testing and deployment for AWS Glue and Lambda jobs
cat full-post.md →--- date: Feb 7, 2025 · slug: spark-application-lifecycle-outside ---
In this blog, we will go in-depth on the overall life cycle of Spark Applications from outside the actual Spark code. Before going ahead, I recommend reading the [Execution Modes]
cat full-post.md →~/.env — vim contact.envchmod 644
1
# contact — portfolio of Vibhavari Bellutagi
3
GITHUB=github.com/vibhabellutagi19
4
LINKEDIN=linkedin.com/in/vibhavari-bellutagi
5
ABOUT="data platforms, pipelines, backend systems, technical writing"