BERLIN, GERMANY / MSc DATA SCIENCE

Hi, I’m

ChaithanyaVemuri.

DATA ENGINEER · ML ENGINEER

I built a restartable pipeline for 411.7M shopping events and verified API delivery and rollback in a Terraform-provisioned GKE lab.

View case studies Download resume GitHub source

Open to junior Data Engineer and ML Engineer roles

INDEPENDENT ENGINEERING & MSc RESEARCHSelected work

GCP / GKE

Verified API delivery and rollbackFulfillAI platform lab

411.7M

Historical events processedREES46 data pipeline

13.93%

Weekly baseline mean CV WAPEOTTO MSc dissertation

01 / SELECTED WORK

Project case studies

A GKE release that couldn’t pull its image, a pipeline that ran out of disk, and a weekly baseline that beat a Transformer.

FULFILLAI / DELIVERY ARCHITECTURE01

{ }SourceGitHub

→

✓Build & scanActions · WIF

→

◇ImageExact digest

GIT-BACKED DESIRED STATE↓

GKE AutopilotArgo CD reconciles the Helm workload

API LAB

Health → metrics → alerts → recovery

01 / FulfillAIDATA & ML · CLOUD PLATFORM LAB

Shipping and recovering an API on GKE

I used Terraform, keyless GitHub Actions and Argo CD to deliver a FastAPI service on GKE. I diagnosed an intentional image-pull failure and verified rollback.

PostgreSQL / dbt
GCP / GKE
Terraform
Argo CD

50,000synthetic orders in the data workflow

Read the case study Source code

REES46 / DATA ARCHITECTURE02

Bronze411,709,736 historical events

RAW → PARQUET

Silver410,325,314 canonical events

VALIDATE / DEDUP

Monthly GoldPartitioned, restartable outputs

FEATURES / FACTS

Then: retrieval → ranking → local API

02 / REES46 V2DATA ENGINEERING · RECOMMENDATIONS

Making a 411.7M-event pipeline restartable

My first aggregation ran out of temporary storage. I split the work by month and session bucket so I could reuse finished work after a retry.

Python
DuckDB
Parquet
FastAPI

411.7Mhistorical events processed

Read the case study Source code

OTTO / SELECTED CV RESULTS03

Selected on CV, before the final holdout

03 / OTTO forecastingMSc RESEARCH · TIME-SERIES FORECASTING

Why a weekly baseline won on OTTO

I compared eight forecasting approaches on four weeks of hourly cart and order counts. Repeating last week’s pattern gave the lowest mean WAPE on the development folds.

Python
PyTorch
scikit-learn
Time-series CV

8approaches compared chronologically

Read the case study Source code

Explore all case studies

02 / HOW I WORK

How I work

What changed after I ran out of disk, compared forecast errors, and tested a bad release.

Check the data first

In REES46, splitting the work only helped if every event in a session still went to the same bucket.

Pipelines & data quality

Keep a baseline

For OTTO, I chose the model on chronological development folds and kept the last 96 hours for the final test.

Evaluation & reproducibility

Test the rollback

In FulfillAI, I checked how Kubernetes behaved when it couldn’t pull the new image and how to get back to a healthy release.

Delivery & operations labs

03 / BACKGROUND

From electronics to data engineering

I studied Electronics & Communication Engineering before moving into Data Science. I’m now pursuing an MSc at Gisma University of Applied Sciences.

Most of my technical work so far comes from independent projects and my dissertation. The case studies link to the code and the experiment records.

More about my background

OPEN TO ENGINEERING ROLES

Get in touch

I’m interested in junior Data Engineer and ML Engineer roles. I’d be glad to discuss the projects or an engineering opportunity.

Contact me