Grammarly Interns Mirror on Their Expertise

Grammarly’s software program engineering internship program is designed to domesticate the following era of tech expertise. Our program combines sensible, hands-on expertise with mentorship, equipping interns with worthwhile abilities for his or her future careers. This yr, we welcomed 12 interns from Germany and the US, with every intern contributing to significant firm initiatives. On this publish, we’ll introduce you to a few of our interns and their tasks and listen to what they discovered from their internship expertise.

Jo Griggs

Staff: Information: High quality and Belief

College: Stanford College, California

Jo labored on an inner instrument known as Stereoscope that permits workers to check textual content suggestions from completely different variations of Grammarly. He built-in giant language mannequin (LLM) suggestions into the instrument by constructing a brand new part that displayed human analysis metrics, LLM analysis metrics, and a comparability between the human and LLM metrics. The undertaking required navigating a number of backends and studying new languages like Clojure. This required Jo to take his technical and mushy abilities to the following stage. For instance, he discovered to interrupt down new codebases and collaborate asynchronously throughout time zones.

To cite Jo: “Having a starvation to study is important to succeed as an intern at Grammarly. I’ve loved this studying course of as a result of I approached it with a want so as to add extra instruments and abilities to my repertoire.”

Khue Nguyen

Staff: Machine Studying Operations

College: Humboldt Faculty, Berlin

Khue was notably excited to hitch the machine studying operations workforce. “With the emergence of huge language fashions, I’m desperate to discover this know-how and see how Grammarly’s product adapts to those new capabilities,” Khue defined. She channeled this ardour into her internship undertaking, during which she assessed varied deployment strategies by benchmarking the efficiency of varied engines and serving frameworks.

Whereas most of her work leveraged Grammarly’s cloud infrastructure, she encountered surprising challenges with a selected vendor framework. After a number of bouts of trial and error didn’t reveal the reply, she puzzled if it may very well be as a consequence of a deeper nuance of the framework design. A gathering with the seller confirmed her speculation, making a path ahead on the difficulty.

Says Khue, “By means of this problem, I’ve discovered concerning the complexity of real-world engineering, the place issues can come up from surprising sources. Because of my workforce, I’ve discovered that efficient problem-solving usually requires trying past the apparent, collaborating with specialists, and pondering creatively.”

Gabriele Matiddi

Staff: Rewrites

College: Sapienza College of Rome

When Gabriele obtained a proposal for a Grammarly internship, it was a no brainer. “I used to be excited by the possibility to work on cutting-edge AI know-how revolutionizing communication,” he defined. “Having used Grammarly for my writing, I used to be desperate to contribute to its growth. I additionally deeply resonated with their values, equivalent to [the importance of being] gritty,” he defined. Throughout his internship, he developed a prototype for voice-based message composition, which challenged him to discover new applied sciences equivalent to giant language fashions and deployment instruments like Docker and Kubernetes. Regardless of his preliminary lack of expertise in these areas, Gabriele’s problem-solving abilities and the help of his workforce helped him succeed.

“Grammarly’s dedication to fostering a various and inclusive atmosphere creates an area the place everybody’s concepts and views are valued,” Gabriele defined. “This made me really feel comfy taking dangers and making an attempt new approaches. For my undertaking, as a substitute of studying each new know-how, I dove into the codebase, experimenting and studying as I went. Every time I hit a roadblock, I may at all times depend on my workforce for steering.” By adopting a learn-by-doing strategy and with the help of his workforce, he was capable of begin contributing on day one.

Yerassyl Altay

Staff: Safety Automation

College: IU Internationale Hochschule, Germany

Yerassyl labored on an initiative to enhance the effectivity of the Removate bot, an inner instrument for mechanically updating code dependencies. His undertaking’s aim was to refine the bot’s scheduling, making certain well timed updates with out disruptions. By means of this undertaking, he deepened his technical understanding of infrastructure, automation tooling, and pipeline administration. Past the technical abilities he discovered, Yerassyl valued the teachings he gained from collaborative discussions with colleagues.

“My workforce fostered constructive debate, viewing disagreements as alternatives for progress,” Yerassyl mentioned. “As an example, they suggested me to prioritize fixing the present downside slightly than preemptively fixing future issues that will not materialize. This mindset has already helped me save time on new tasks.”

Artem Yushko

Staff: Core Product

College: Carleton Faculty, Northfield, Minnesota

Artem contributed to a undertaking that sought to grasp the connection between a author and their viewers. To research the info, Artem wrote a service from scratch—an endpoint that pulled and filtered knowledge throughout a number of sources. Given the complexity of the present codebase, the duty was demanding. Nonetheless, Artem succeeded by collaborating intently along with his mentor and workforce on the undertaking scope and technical strategy.

“Because of the mentorship and help from my supervisor and mentor, I may study the very best practices instantly with out stumbling my manner by,” Artem mentioned. By means of the collaboration, Artem discovered that efficient teamwork is essential to delivering high-quality code.

Johan City

Staff: Development, Paid Conversion

College: Radboud Universiteit, Nijmegen, Netherlands

Johan spearheaded an experiment providing superior strategies to new Grammarly customers, showcasing the worth of those premium Grammarly options. This cross-functional undertaking required shut collaboration with knowledge science, design, and product administration. By means of this collaboration, Johan discovered how engineering may act as a bridge between teams with completely different contexts and views.

“Within the preliminary phases, I proactively communicated the product’s technical constraints, enabling us to adapt the scope with out compromising our capability to study,” he defined. All that arduous work paid off: The experiment is now dwell, offering worthwhile insights even after Johan’s internship has concluded.

Trying ahead

With all this unimaginable studying, we’re excited to see what our interns do subsequent. If you wish to comply with of their footsteps, we encourage you to use for subsequent yr’s internship program.