April 19, 2024


Ever puzzled what it’s wish to intern as a software program engineer at Slack? Image your self on the well-known Ohana ground—the 61st ground of the Salesforce Tower in San Francisco— it’s one among many privileges we had as interns. Not solely did our expertise with Slack’s Information Engineering group allow us to step onto the tech frontier, however it additionally gave us the prospect to realize invaluable engineering expertise and forge relationships throughout groups. Most significantly, we bought to contribute to Slack’s mission by means of two impactful tasks: The Dependable Information Discovery Instrument helped customers uncover related information property, whereas the Job Efficiency Monitoring and Alerting system empowered engineers with insights into end-to-end efficiency metrics. Persist with us to listen to  about what it was wish to be a Information Engineering intern on varied tasks, and what made us profitable on this journey.


The onboarding expertise for all Information Engineering interns was a seamless and well-orchestrated course of, accommodating our various begin occasions. We had been divided into two distinct two-week parts, the primary targeted on Salesforce and Slack integration. We participated in a collection of informative conferences, partaking break-out classes, and hands-on tutorial prep classes. This complete strategy left us interns feeling well-prepared and assured as we launched into our internship journey.

Following this, the second portion encompassed per week solely devoted to onboarding onto the Information Engineering pillar. Throughout this time, we had been supplied with detailed setup documentation, guaranteeing we had the mandatory instruments and assets at our disposal. Extra importantly, this week emphasised team-building, fostering a welcoming and useful setting. We had the chance to satisfy and get to know everybody on the group, permitting us to forge connections and really feel like an integral a part of the Information Engineering group at Slack from day one.

Double the tasks, double the affect!

As Information Engineering interns, we helped advance Slack’s mission by means of two main tasks, spanning throughout 4 core groups at Information Engineering. The Dependable Information Discovery Instrument helped customers with a greater information search expertise, whereas the Job Efficiency Monitoring and Alerting system enabled focused ETL job efficiency optimization for information engineers.

Dependable Information Discovery Instrument

Camryn McDonald, Esha Thota

Downside assertion: Our inside analytics instrument permits Slack staff to look dashboards (1,700), queries (1.8 million), and tables (11,000) to retrieve information. That’s a variety of information! Because the numbers of every of those present search gadgets enhance, the information search/retrieval course of turns into more and more gradual and unreliable. How can customers really know what makes a very good dashboard or desk? How can we make it straightforward for customers to find the data related to their use instances?

Customers are lacking an easily-navigable, dependable search expertise that makes it easy for them to find the information they want and achieve perception into its reliability and usefulness. Through the use of our backend resolution of OpenSearch — an open-source search and analytics platform constructed on Elasticsearch — and making a frontend that makes our information straightforward to entry and parse, we’re capable of help our customers higher to find the information that they want.

Backend: Diving into backend work with this intern venture, we utilized OpenSearch — an improved resolution in comparison with our present implementation. We labored to index and search metadata from DE’s inside information property, return ranked search outcomes for most important object varieties, and show related metadata alongside outcomes to point reliability. All of this work helped to handle the difficulty of scalability as we proceed to get an increasing number of information.

Frontend: All through the course of our intern venture, we really bought to take a deep dive into design. Using Typescript, React, and AntDesign, we constructed out a versatile and simply navigable search UI for customers to search out, type, and filter their information. It might probably return related details about information objects (together with title, proprietor, month-to-month energetic customers and warnings, and so forth.), and manages the state of ends in URL for simpler sharing of findings. We discovered about what really makes a very good person experience- and the right way to design for it!

Integration and Downside Fixing: Having labored on completely different points of the venture, we gained expertise in integrating varied frontend options seamlessly with the backend, using the Quick API. We discovered to work collectively to strategy difficult issues, the right way to take care of blockers, and the right way to ask the precise inquiries to get the solutions we wanted.

Job Efficiency Monitoring and Alerting

Donavon Clay, Tayven Taylor

Downside assertion: Throughout a lot of Slack’s engineering groups, hundreds of knowledge pipelines run, course of, and generate petabytes of knowledge day by day (yeah, that’s lots). Generally, these processing jobs fail due to efficiency points, and groups must take the time to research the failure and take corrective motion. Whereas this course of works, it’s inefficient and takes our engineers’ time away from different tasks. What groups are lacking is a complete view into historic job efficiency that offers them the instruments to handle their pipelines proactively, as an alternative of reactively.

Our venture takes information from two most important sources, the Spark Historical past Server and Apache Airflow, and creates an all-encompassing desk with varied metrics that’s used to visualise groups’ historic job efficiency on inside dashboards. With the flexibility to have a look at how their jobs carry out over time, we give groups the chance to optimize their jobs and useful resource utilization, whereas limiting failure.

Spark metrics: This neat UI generally known as the Spark Historical past Server permits us to dissect our information processing jobs and have a look at efficiency metrics (assume reminiscence utilization, runtime, and so forth.). After working by means of some documentation, we discovered the right way to arrange an ingestion job that queried the server’s API, parsed the efficiency information, and saved it for later utilization. There was undoubtedly a studying curve to it, however we had been capable of achieve some worthwhile data of instruments utilized in information engineering.

Airflow metrics: Airflow is the instrument we use to schedule batch jobs. At a excessive degree it capabilities kind of like an ordered to-do record (do that process, then as soon as it’s completed, do that). The information we had been desirous about pulling from right here included issues like group possession and efficiency expectations. We spent some high quality time with the documentation to get the cling of the know-how after which arrange ingestion pipelines to tug in all the information and fields we wanted, doing sanity checks to verify it labored correctly.

Enriched desk and dashboards: After engaged on each Spark Metrics and Airflow Metrics datasets, the ultimate step was to merge them collectively to get a desk that has not solely job efficiency, however group identification and configurations as properly. Utilizing this enriched desk, we created dashboards for groups to have a look at their most computationally costly duties, historic job developments, and arrange alerts to sign when a job would possibly fail. This was one thing that hadn’t beforehand existed, so many groups had been excited to see it. We had been glad to see we made an affect, all while studying these applied sciences we hadn’t had an opportunity to see in class.

The TPM perspective

Naman Modani

As a Technical Product Administration (TPM) intern on the Information Engineering group, I had the distinctive alternative to work cross-functionally on a number of high-impact tasks. Past managing the supply of the Dependable Information Discovery Instrument and Job Efficiency Monitoring and Alerting initiatives, I additionally helped drive two concurrent flagship packages: Information Retention, aligning with Slack’s privateness objectives, and Legacy Customers Deprecation, changing outdated, much less environment friendly tables with quicker, modular ones.

My function concerned aligning the product and engineering imaginative and prescient, facilitating clear communication between stakeholders, and proactively surfacing dangers, bottlenecks, and dependencies. With the assistance of my mentor and supervisor, I discovered a lot about what it takes to efficiently handle tasks in an agile setting. Whereas no day was the identical, my key takeaways embody:

  • Figuring out and aligning on scope early is essential. This enabled us to interrupt down epics into actionable tales whereas mitigating dangers upfront.
  • Efficient stakeholder administration is mission-critical. I strived to maintain all events constantly looped in by means of conferences, documentation, and weekly updates. This open movement of knowledge enabled seamless collaboration, and helped drive accountability by means of energetic listening and engagement with Product, Engineering, and management.
  • Flexibility and adaptation are important. Necessities evolve quickly, so I needed to stay nimble and modify timelines accordingly. Agile rules stored me grounded when plans modified.
  • Technical acumen goes a good distance. Actively growing foundational information/methods data helped me higher contextualize tasks and converse knowledgeably with cross-functional companions. This additionally helped me automate guide processes for a number of packages I labored with.
  • Comfortable expertise make all of the distinction. Endurance and empathy had been instrumental in working by means of advanced technical challenges and constructing sturdy intra-team connections. The human ingredient was invaluable.

I cherished diving deep into Slack’s information methods and serving to drive significant initiatives. Cross-functional work gave me worthwhile insights into delivering impactful tasks amidst fast change. I’m thrilled to use these learnings in my future roles.



As we mirror on our internship expertise—particularly on what made our intern journey profitable—a couple of key themes emerge:

  • The chance to collaborate on real-world tasks enabled impactful hands-on studying. Tackling advanced challenges allow us to achieve sensible engineering expertise past textbook data. Working cross-functionally with our gifted mentors and managers helped us obtain greater than we might have individually.
  • The unwavering help of our mentors empowered our development. Every time we felt caught, they patiently guided us by means of technical roadblocks. Their insights and encouragement pushed us to continue learning and bettering. By working intently with our mentors, we grew tremendously each technically and personally.
  • The welcoming and enjoyable Slack tradition made every day fulfilling. Bonding throughout actions like workplace onsites (minigolf!), and lunch outings fostered camaraderie. This helped us construct significant connections throughout groups and considerably improved group collaboration.
  • Navigating challenges developed our resilience and adaptableness. From relocating to new cities to tackling advanced information methods, we discovered to embrace change and push previous our consolation zones. These experiences helped us mature professionally and personally and laid a strong basis for our future profession.

Total, these elements helped make our internship a hit at Slack. We’re deeply grateful for the teachings, connections, and reminiscences from our time right here. As we every embark on the following chapters in our careers, we’ll carry the inspirational Slack ethos with us.