Chandrapal, data engineer

Chandrapal

Senior Data Engineer

Wrocław, Poland · Europe/Warsaw (UTC+1) · 11 years of experience

Chandrapal focuses on dependable data products with tested transformations, lineage, and freshness controls.

Request a shortlist

Why this profile fits

Trustworthy data products

Defines contracts, tests, ownership, lineage, and freshness expectations for data used by teams and models.

Scalable processing

Chooses batch, streaming, and storage patterns around volume, latency, recovery, and operating cost.

Operable pipelines

Designs idempotent jobs, backfills, alerting, and incident procedures rather than treating a successful first run as done.

Relevant toolkit

SQLPythondbtSparkAirflow

Delivery evidence

Selected work

Uplift-Based Retention Engine

Subscription streaming platform

Replaced a blanket save-offer program with a two-model uplift framework that targeted only persuadable subscribers rather than everyone showing risk. Cut retention discount spend by 31% while lifting treated-cohort save rate by 4.2 points, validated through a 6-week randomized holdout before any rollout.

Pythonscikit-learn

Churn Early-Warning Scoring Pipeline

Digital telecom operator

Built a daily survival-analysis pipeline scoring tens of millions of prepaid and postpaid subscribers for lapse risk. Predictions fed CRM journeys and outbound call lists, reducing 90-day voluntary churn by 18% in the highest-risk decile within two quarters.

Pythonlifelines
View more profile detail →

Relevant experience

Career timeline

Senior Data Scientist, Retention & Lifecycle

Mar 2021–Present

Devlyn (remote client engagements) · Wrocław, Poland (Remote)

  • Led churn and retention modeling for a rotating portfolio of subscription, fintech, and streaming clients, shipping uplift and survival models that lifted net revenue retention by 3 to 6 points per engagement.

System Administrator

Jul 2017–Present

ViitorCloud Technologies · IND, On-site

  • Owned the subscriber churn model for a catalog serving 3.4 million active subscribers, moving the team from a single classifier to an uplift model that treated only persuadable users.

Data Scientist

Aug 2016–May 2018

Digital telecom operator · Warsaw, Poland

  • Developed survival-analysis churn scoring for tens of millions of prepaid and postpaid subscribers, reducing 90-day voluntary churn 18% in the highest-risk decile.

Skill depth

Show experience in context.

VerbalCommunicationDomainUnderstandingProblemSolvingCodeQualitySystemDesignDeliverySpeed
Competency shapeAssessment across the same six dimensions used for every profile.

Relevant experience by skill

Python11 years
SQL11 years
scikit-learn10 years
pandas10 years
Churn & Retention Modeling9 years
A/B Testing & Experimentation8 years
PyTorch7 years
Snowflake6 years
See the complete skill index

Languages

PythonSQLRScalaBash

Libraries/APIs

pandasPolarsscikit-learnstatsmodelsXGBoostLightGBMPyTorchTensorFlowlifelinesSHAPOptunaLangChain

Frameworks

dbtAirflowDagsterMetaflowFastAPIStreamlitRay

Tools

MLflowWeights & BiasesGreat ExpectationsDockerGitJupyterTableauLooker

Paradigms

Survival AnalysisUplift ModelingA/B TestingCausal InferenceBayesian InferenceTime-Series ForecastingFeature EngineeringMLOps

Platforms

SnowflakeBigQueryDatabricksAWS SageMakerGCP Vertex AIKubernetes

Storage

SnowflakeBigQueryPostgreSQLpgvectorRedisDelta LakeParquetAmazon S3

Other

Net Revenue RetentionLifetime Value ModelingModel Monitoring & Drift DetectionReactivation & Win-back Analytics

Education

Formal background

Master of Science, Mathematics, Statistics specialization

University of Wrocław · 2013–2015 · Graduated with distinction

Bachelor of Science, Computer Science

Wrocław University of Science and Technology · 2010–2013

Credentials

Certifications

Databricks Certified Data Engineer Associate

Databricks · 2023 · Certified

dbt Analytics Engineering Certification

dbt Labs · 2022 · Certified

Communication

Languages

Polish

Native

English

Fluent

German

Conversational

More examples

Similar Data Engineer profiles.