<div class="show-more-less-html__markup show-more-less-html__markup--clamp-after-5
 relative overflow-hidden">
 What you bring Must Haves <ul>
<li>3-4 years in Machine Learning roles, building, prototyping and deploying language models</li>
<li>Proficiency with core ML and LLM tech stack: Python ecosystem, Vector Databases (Weaviate, Pinecone) Tensorflow/Pytorch, Huggingface, Langchain, AWS (SageMaker, etc.).</li>
<li>Deep understanding of LLM concepts, including prompt engineering, transformers, reinforcement learning from human feedback (RLHF), model fine-tuning, and embeddings.</li>
<li>Experience with retrieval, recommendation and search algorithms 
</li>
</ul>
Nice To Haves <ul>
<li>Have published papers at ML/NLP conferences like ACL, EMNLP, CoNLL, ICML, RecSys, etc.</li>
<li>Experience of working with data processing tools at scale (Python, Spark, Hadoop, etc. )</li>
<li>Knowledge of working with containerized cloud environments (Docker, Kubernetes, AWS/GCP) 
</li>
</ul>
 </div>

Software Engineering

(Other)

(Engineering)

Costa Rica

We built a hypercloud with bare-metal performance. A hypercloud is a 100% software configurable virtual datacenter that lives on top of physical hardware securely in your private datacenter, on the public clouds of AWS, GCE, and Azure, and on your laptop. Virtual Machines tax performance. Containers are too new and insecure, forcing you to redesign app deployment flow. The hypercloud unit of computation is the Terminal Machine (TM, or just Terminal). TMs are the best of both worlds: you get the security and familiarity of full operating system instances with the bare-metal speed and efficiency of containers. TMs rebalance and migrate seamlessly without reboot to maximize efficiency of your cluster. Terminals resize from 1 CPU to 32 CPUs and 50GB of RAM when you need the power, in seconds, with no reboot. They automatically pause or shrink back down to save you money. “The software-defined data center is 10 years away,” they say. “The public cloud is too insecure,” they say. The hypercloud is the best of both worlds: you get the abstraction and power of the cloud, with the security of running everything in your own datacenter, plus the benefits of a hybrid cloud if you want both. Today. Who uses Terminal.com? Our users range from web app developers who spin up copies of their website in 5 seconds for QA and testing, to neuroscientists who analyze large FMRI datasets on computers that scale instantly, to online schools who teach people how to code using Terminal.com. RailsBridge relies on Terminal.com to teach first-time female engineers to code Ruby on Rails. Top consultants at ZS Associates crunch numbers on our platform for data science tasks every day. Professors at Carnegie-Mellon use Terminal.com to teach 100-student courses with no wasted effort. Engineers across Silicon Valley know that Terminal.com is the easiest and most powerful tool for conducting coding interviews. Terminal.com provides the full suite of tools for the modern developer. How did Terminal.com start? When Varun Ganapathi’s family moved from India to Boston, Massachusetts in the 1980's, they started a small company called Terminal Exchange Systems. The company provided terminals to local companies so that they could remotely access huge, main-frame computers. Varun Ganapathi helped his father carry machines door to door, renting out both the hardware and software to local businesses. As he grew up, Varun’s commitment to computing didn’t disappear. He completed his computer science PhD at Stanford, sold over 1 million copies of his paid photography mobile app, Pro HDR, and sold his second company, Numovis, to Google. After two years as a Google Senior Research Scientist, Varun asked his father if he could use the company domain name, Terminal.com. Varun explained that he had plans for it, but this time there would be no need to carry hardware door to door. Today, Varun and his San Francisco-based team rent out browser-based virtual machines that they call Terminals to thousands of users across the globe giving them access to the infinite resources of the cloud. Users ranging from online education students to Dev-Ops professionals can now start, resize, and scale their own Terminals in a matter of seconds. Inside each Terminal, they discover an intuitive interface for creating, editing, and sharing programs. And they don’t even have to wait for them to be delivered to their door.

San Francisco, CA, USA

Terminal.com

Machine Learning Engineer

Craft Ventures

Tell us about your professional DNA to get discovered by any company in our network with opportunities relevant to your career goals.

Leverage our network to build your career.

#### Welcome!

Thank you for joining Craft Ventures Network!

To help us best support you in your search, please take a few minutes to tell us about what you are looking for in your next role. We’ll use this information to connect you to relevant opportunities in the Craft Ventures network as they come up.

You can always update this information later.

As our companies grow, they look to us to help them find the best talent.

Signal that you'd be interested in working with a Craft Ventures company to help us put the right opportunities at great companies on your radar. The choice to pursue a new career move is then up to you.

Opportunities at Craft portfolio companies

Machine Learning Engineer