Citi Data Engineer interview questions & guide 2026

Interview evaluations at Citi cover technical concepts, distributed system mechanics, and hands-on coding. These questions are drawn directly from real candidate experiences across global locations and varying team functions. The goal is to evaluate your understanding of underlying computing engines, data modeling, and practical query optimization rather than simple syntax memorization.

PySpark & Distributed Computing

This category tests your deep understanding of Apache Spark execution plans, internal memory management, data partitioning, and job optimization. Candidates face both theoretical questions on framework mechanics and practical scripting tasks.

• Explain the difference between repartition() and coalesce() in PySpark. Which performs better in reducing output file count and why?
• How do you diagnose and handle severe data skew in a Spark cluster when dealing with skewed join keys?

Preparation for Citi requires a dual focus on practical hands-on coding and deep system mechanics. Interviewers probe whether you truly understand how distributed framework APIs execute under the hood, how operating systems manage data jobs, and how data schemas scale across enterprise architectures.

Role-Related Knowledge – Candidates must possess strong command over PySpark, SQL, distributed computing internals, and transactional data processing. Evaluation focuses on your ability to write clean pipelines, choose optimal file formats (Parquet vs. Avro), and tune resource parameters. You demonstrate strength by detailing execution plans, memory allocation, and frame optimization.

Problem-Solving & System Diagnostics – When data platforms fail or slow down, Citi expects engineers to methodically diagnose root causes. Interviewers present scenario-based questions involving stuck Spark jobs, memory overflow errors, and dataset migrations. You can showcase competency by explaining systematic isolation procedures, analyzing job UI metrics, and applying targeted mitigation strategies.

Enterprise Data Architecture & Governance – Designing pipelines for a bank requires strict attention to data validation, schema evolution, and access patterns. Candidates must articulate clear strategies for partitioning, bucketing, SCD Type 2 tracking, and metadata verification. Highlighting automated quality framework integration and data governance standards builds immediate credibility.

The hiring process for a Data Engineer at Citi typically spans 3 to 5 weeks, moving from basic technical screens to in-depth technical panel discussions. The experience balances core algorithmic or scripting competency with real-world infrastructure scenario questions.

Depending on the team and location, your initial technical screening may be conducted via a standardized assessment partner like Karat or directly by a Citi engineering manager. Initial rounds evaluate database principles, SDLC practices, standard Python operations, and core query formulation. Be prepared for direct script-writing and code output prediction during these early screens.

Subsequent technical rounds transition into comprehensive face-to-face or virtual panel interviews. These sessions focus on big data architecture, PySpark execution tuning, migration patterns (Teradata or Oozie to Spark and Kafka), shell command syntax, and data partitioning design. Candidates frequently note that Citi delves deeper into on-premise Hadoop and open-source engine internals than many traditional cloud-first tech companies.

The timeline above illustrates the standard progression from initial recruiter intake through specialized technical evaluations to final leadership alignment. Candidates should pace their prep by building foundational strength in SQL and Python first, followed by intense review of big data execution internals before panel interviews. While exact stages vary slightly across global hubs, technical rigor remains consistent.

PySpark & Big Data Processing

This evaluation area tests your mastery over PySpark execution engines, job optimization techniques, and memory management across distributed worker nodes. Interviewers want to verify that you understand how code transformations map to physical cluster operations.

Be ready to go over:

• Spark Execution Architecture – Understanding drivers, executors, tasks, stages, narrow vs. wide transformations, and physical execution plans.
• Data Skew Diagnostics – Identifying key-skew bottlenecks during joins and applying salting or broadcast strategies.

As a Data Engineer at Citi, your day-to-day work centers around designing, building, and maintaining robust data platforms that process critical enterprise data. You will work across the full software development lifecycle, transforming raw data requirements into operational, low-latency data pipelines.

A primary deliverable involves constructing robust ETL/ELT pipelines using PySpark, Databricks, Python, and standard relational databases. You will be actively involved in modernization efforts, migrating legacy codebases from systems like Teradata or Oozie to distributed frameworks using Spark, Kafka, and modern cloud object stores. Protecting historical integrity through SCD Type 2 tracking and enforcing strict validation checks using automated tools form core components of your workflow.

Collaboration is an essential aspect of the job. You will work alongside quantitative analysts, risk modeling teams, business analysts, and compliance officers to define precise data contracts and schema definitions. Additionally, you will coordinate with enterprise infrastructure teams to optimize resource usage on large Hadoop clusters or cloud compute engines, maintaining optimal performance while minimizing operational overhead.

Candidates applying for Data Engineer positions at Citi must balance deep core software development skills with enterprise big data experience. Requirements vary across individual titles (ranging from Senior Analyst up to Vice President levels), but foundational expectations remain highly consistent.

Technical Skills

• Distributed Systems & Big Data: Strong proficiency in PySpark, Apache Spark architecture, Hive, and distributed processing fundamentals.
• Programming Languages: Fluency in Python and SQL is mandatory. Working knowledge of Scala or Shell scripting is highly desirable.
• Data Warehousing & Modeling: Demonstrated mastery of enterprise dimensional modeling, SCD Type 2 pattern implementations, and file format mechanics (Parquet, Avro, ORC).
• Data Quality & Governance: Familiarity with validation frameworks such as Amazon Deequ, data profiling tools, and enterprise catalog systems.

Experience & Soft Skills

• Prior Experience: Typically 3+ years of dedicated data engineering experience for individual contributor roles, with senior roles requiring 7+ years managing enterprise-scale data infrastructure.
• Operational & Analytical Mindset: Strong ability to debug complex distributed job failures, analyze performance bottlenecks, and tune system queries.
• Communication & Stakeholder Management: Ability to explain technical trade-offs, schema structures, and pipeline limitations clearly to non-technical business partners and audit risk panels.

Must-Have Qualifications:
- Advanced SQL proficiency (CTEs, analytical functions, execution plan analysis)
- Hands-on PySpark engineering experience in production environments
- Proficiency in Linux/Unix shell scripting and background task management
- Core understanding of relational and big data modeling patterns

Nice-to-Have Qualifications:
- Experience with Apache Iceberg, Starburst/Trino, or Databricks platforms
- Hands-on streaming experience using Apache Kafka or Spark Streaming
- Exposure to legacy technology migration (Teradata, Oozie, on-prem Hadoop)

Q: How technical are the initial screening rounds? Initial screens, often conducted via Karat or virtual assessment panels, are highly technical. Expect direct Python scripting tasks, output prediction challenges, standard SQL query challenges, and rapid-fire theoretical questions covering database management systems and SDLC principles.

Q: Does Citi focus more on cloud platforms or on-premise technologies? While Citi is actively expanding its cloud presence using platforms like Databricks and Snowflake, many core financial engines still rely on legacy on-premise Hadoop, Hive, Oozie, and Teradata systems. Be prepared to answer questions covering both modern cloud patterns and on-premise big data mechanics.

Q: What is the primary difference between a Data Engineer and a Big Data Lead at Citi? Mid-level Data Engineers focus heavily on pipeline script creation, query writing, and basic optimization tasks. Lead or VP-level candidates are evaluated on overall platform architecture, cross-system migration strategies (SCD Type 2 handling), cluster resource management, and team leadership.

Q: How much coding vs. theory should I expect in the final panel interviews? Final panels balance theoretical big data architecture with practical live coding. Expect roughly equal weight given to conceptual questions (repartition vs. coalesce, file format trade-offs) and hands-on coding (writing complex SQL queries or PySpark transformations).

• Structure Your Coding Answers: When asked to write complex SQL or PySpark scripts, clearly state your schema assumptions, step through your approach using modular logic (such as Common Table Expressions), and explain your indexing or partitioning strategy before typing.
• Highlight Optimization Metrics: In financial engineering, runtime efficiency matters. Always discuss how your design minimizes network shuffling, reduces memory pressure, and avoids unnecessary disk I/O.
• Be Prepared for Syntax Integrity: Unlike interviews that allow pseudocode, Citi interviewers frequently request clean, executable PySpark transformations or standard shell commands during face-to-face or virtual coding rounds.

• Detail Data Integrity & Quality Mechanisms: Highlight how you ensure data accuracy. Mention explicit data validation frameworks (Amazon Deequ) and standard schema constraint checks in every pipeline design scenario.

Targeting a Data Engineer position at Citi places you at the threshold of solving complex big data challenges across global financial systems. The role demands technical rigor, strong software engineering foundations, deep distributed computing expertise, and a comprehensive commitment to data quality and operational stability.

To maximize your interview performance, focus your preparation on core execution mechanics. Practice writing clean, performant PySpark transformations, master complex SQL analytical query patterns, review Linux operational commands, and be prepared to articulate exact diagnostic strategies for stuck or failing batch pipelines. Demonstrating clear logic and system-level awareness will distinguish you as a strong candidate.

The compensation data above illustrates base salary expectations for Data Engineering roles across geographic regions and seniority bands. When evaluating these numbers, consider total compensation structures, including performance bonuses and role seniority levels within Citi's corporate hierarchy.

For candidates seeking deeper interview prep resources, real candidate interview write-ups, targeted practice problems, and platform insights, explore comprehensive prep materials on Dataford to refine your technical readiness and approach your interviews with confidence.