← Back to Directory

Databricks interview notes, page 10 of 22

Large Enterprises

Most of these Databricks write-ups come from Software Engineers (192), then Machine Learning Engineers (9) and Data Scientists (8). The most reported rounds are the technical phone screen (136), the system design round (57), the behavioral / leadership round (9) and the onsite coding round (6).

216 interview notes · updated Jul 2026

Databricks·Software Engineer·Technical Phone Screen

May 2026
Databricks SWE interview that went pretty deep into streaming systems. Part one was manageable but the follow-up on memory-bounded RLE caught me off guard in a good way, really made me think about what 'correctness' even means when you're flushing state.
  • Implement run-length encoding over a stream of tokens, outputting (value, run_length) pairs for consecutive equal tokens. Your implementation must also handle backpressure from a slow downstream consumer.
  • Now imagine the token stream is extremely high volume and you can't keep all encoded pairs in memory. How do you redesign the RLE to only retain the last K encoded pairs, emit earlier ones to a downstream sink, and still guarantee correctness at the boundary between flushed and in-memory state?

“The core RLE part was fine, just track current token and count, emit when it changes.”

View Post

Databricks·Software Engineer·Technical Phone Screen

May 2026
Databricks SWE interview with a two-part coding problem centered on networking fundamentals. The IP validation part was manageable but the CIDR follow-up required some bitwise thinking that I wasn't fully warmed up for.
  • Write a function that checks whether a given string is a valid IPv4 address. Each octet must be between 0 and 255, and leading zeros should be rejected.
  • Given an IP address and a CIDR block like 192.168.1.0/24, determine whether the IP falls within that CIDR range.

“Felt pretty solid on this one.”

View Post

Databricks·Software Engineer·Technical Phone Screen

May 2026
Databricks coding round focused on a compression problem that combined two encoding schemes. The encoding direction was the real puzzle, and I spent most of my time second-guessing the decision logic for which scheme to apply.
  • Implement a compress and decompress function that applies Run-Length Encoding when there are at least 8 consecutive repeated values, and Bit-Packing when exactly 8 values can be packed together. The order of values must be preserved, and you have to decide on the fly which scheme to apply to each segment.

“The decompression part is almost mechanical once you understand the format your encoder produces.”

View Post

Databricks·Software Engineer·Technical Phone Screen

May 2026
Databricks SWE interview that was more systems-flavored than I expected. The main problem was about QPS tracking for a key-value store, but the real meat was in the follow-ups, which kept pushing toward smarter tradeoffs.
  • Given a key-value store, how would you compute the queries per second for put and get operations?
  • How would you support QPS queries over an arbitrary time interval [t1, t2], not just a fixed window?
  • Can you modify the approach to trade some accuracy for reduced memory usage?

“Base question felt manageable.”

View Post

Databricks·Software Engineer·Onsite - System Design / Architecture

May 2026
Databricks system design round focused entirely on a VM allocation service. One question, fairly scoped, but there's more depth hiding in it than you'd expect from the initial prompt.
  • Design a service that allocates VMs by attaching them to free IPv4 addresses from a shared pool. Cover the API surface, database schema for the IP pool and allocation records, and the request workflow for attaching and detaching a VM to an IP.

“My first instinct was to jump straight into the schema and I had to pull myself back.”

View Post

Databricks·Machine Learning Engineer·Technical Phone Screen

May 2026
Databricks ML engineer round focused entirely on implementing a Lazy Array from scratch. Pretty niche for an MLE role but made sense given how much they care about data pipeline efficiency. The follow-up about testing laziness was actually the more interesting part.
  • Implement a Lazy Array data structure that defers computation until elements are actually accessed, supporting map, filter, take, and indexed access without materializing intermediate arrays.
  • How would you verify that your Lazy Array implementation is genuinely lazy and not just appearing to be? Walk through a testing approach.

“I got the basic idea fast enough but fumbled on composing multiple transformations without intermediate allocations.”

View Post

Databricks·Software Engineer·Onsite - System Design / Architecture

May 2026
System design round at Databricks for a software engineering role, focused entirely on building a stock trading system on top of a third-party exchange API. Pretty deep dive, lots of follow-up on failure modes and scalability.
  • Design a stock trading system that supports buy and sell orders (market and limit), built on top of a third-party exchange API with endpoints to post, query, and cancel orders.
  • How would you reliably detect and handle order expiration based on time-in-force constraints?
  • How do you ensure idempotency and handle retries when calling the third-party exchange API?
  • What are the main bottlenecks in this system at scale, and how would you address them?

“I started with the data model and order lifecycle which felt right, but I think I underexplained the state machine early on.”

View Post

Databricks·Software Engineer·Onsite - System Design / Architecture

May 2026
Databricks system design round focused entirely on concurrent data structures. One long question about thread-safe hash maps that went pretty deep into JVM internals. Left feeling like I covered the surface but probably missed some nuance they were looking for.
  • Design and implement a thread-safe hash map that handles many concurrent readers and writers efficiently. Walk through why a plain HashMap fails under concurrency, what coarse-grained alternatives exist, and how ConcurrentHashMap actually works internally.

“I started with the obvious stuff, HashMap isn't thread-safe because concurrent rehashing can corrupt the internal structure, and in older JDKs you could get threads stuck in infinite loops reading a circular linked list.”

View Post

Databricks·Software Engineer·Technical Phone Screen

May 2026
Databricks Solution Architect interview that split into two parts: a project walkthrough and then a pretty deep technical dive into database fundamentals and distributed systems. The technical portion was more rigorous than I expected for an SA role, felt more like an engineering loop.
  • Walk me through your most relevant data engineering or analytics projects.
  • Compare OLAP and OLTP systems. How do workload patterns, schema design choices, and storage engines factor into when you'd pick one over the other?
  • How do you detect data skew in a distributed processing job, and what are the main ways to fix it?

“Pretty standard opener but I rambled a bit trying to cover too many projects.”

View Post

Databricks·Software Engineer·Recruiter / HR Screen

May 2026
HR screen at Databricks, just the one question about why I was looking to leave. Pretty standard stuff but worth thinking through before you go in.
  • Why are you looking to leave your current job?

“You really can't wing this one.”

View Post