Artur Kiulian is an entrepreneur and investor in the field of Artificial intelligence. He is best known as the founder of Colab, a company that developed a platform for AI-powered collaborative workspaces. Kiulian's career spans both corporate research and startup leadership, with a focus on applying Machine learning and Deep learning to practical business problems.
Born in the late 1980s, Kiulian began his professional career in the technology sector in the early 2010s. He initially worked in software engineering and product management roles, gaining experience in the consumer electronics and cloud computing industries. His early work involved developing applications for mobile devices and integrating Neural network models into user-facing products.
Early Career and Corporate Experience
Kiulian's first significant corporate role was at Samsung Electronics, where he worked from 2012 to 2015 as a product manager in the company's mobile division. During this period, he was involved in the development of features that used early Machine learning algorithms for image recognition and voice commands. He contributed to the launch of the Samsung Galaxy S4 in 2013, which included smart-scroll and smart-pause features that relied on eye-tracking technology.
In 2015, Kiulian moved to Apple as a senior software engineer in the Siri team. He worked on improving the natural language processing capabilities of the virtual assistant, focusing on the Sequence-to-Sequence (Seq2Seq) models that were then emerging in Deep learning research. His tenure at Apple lasted until 2017, during which he filed several patents related to voice-activated device controls and context-aware computing.
Transition to Entrepreneurship
In 2017, Kiulian left Apple to co-found his first startup, a company called Omniscient, which specialized in AI-driven data analytics for retail businesses. The company developed a platform that used Transformer (architecture) models to analyze customer purchase patterns and predict inventory needs. Omniscient raised $2.3 million in seed funding from a group of angel investors in 2018, but the venture struggled to gain traction and was acquired by a larger analytics firm in 2019 for an undisclosed amount.
Following the acquisition, Kiulian spent a year as a visiting researcher at Stanford AI Lab, where he collaborated with faculty on projects related to Large language model interpretability. This academic period influenced his subsequent work, particularly his interest in making AI systems more transparent and controllable for end users.
Founding of Colab
In 2020, Kiulian founded Colab, a company aimed at creating a collaborative workspace where teams could interact with AI assistants in real time. The platform integrated Generative AI models, including early versions of OpenAI's GPT-3, to help users draft documents, generate code, and manage project workflows. Colab's key innovation was its focus on multi-user interaction with a single AI instance, allowing teams to share context and build on each other's prompts.
Colab launched its beta version in March 2021 and attracted 50,000 users within the first six months. The company raised a $5 million Series A round in November 2021, led by the venture capital firm Halcyon. By the end of 2022, Colab had expanded its user base to 200,000, with paying customers in the software development and marketing sectors. The platform was noted for its use of Reinforcement Learning from AI Feedback (RLAIF) (reinforcement learning from AI feedback) to fine-tune its models based on user interactions.
Investment and Advisory Roles
In addition to his entrepreneurial work, Kiulian has been an active angel investor since 2019. He has made early-stage investments in several AI startups, including Fermata, a company developing computer vision systems for agriculture, and XYber, a cybersecurity firm that uses anomaly-detection algorithms. Kiulian has also served as an advisor to the AI research lab AI21 Labs, where he provided guidance on product strategy for their Jurassic-1 language model.
In 2022, Kiulian joined the board of directors at Insta Academy, an online education platform that offers courses in Machine learning and data science. He has been involved in developing the curriculum for their advanced deep learning track, which covers topics such as Residual Network (ResNet) architectures and Batch Normalization techniques.
Technical Contributions
Kiulian has published several papers on applied AI, with a focus on efficient model deployment. In 2021, he co-authored a paper on Model Pruning techniques for edge devices, which demonstrated a 40% reduction in inference time for image classification models without significant loss of accuracy. The paper was presented at the International Conference on Learning Representations (ICLR) in 2022.
He has also contributed to open-source projects, including a library for Data Augmentation in natural language processing that is used by several research groups. His work on Curriculum Learning strategies for training Transformer (architecture) models has been cited in academic literature on efficient training methods.
Public Speaking and Media
Kiulian is a frequent speaker at technology conferences, including the AI Summit in London and the NeurIPS industry track. His talks often focus on the practical challenges of deploying Large language models in production environments, such as managing Temperature Scaling and Top-P (Nucleus) Sampling parameters to control output quality. He has also appeared on podcasts and webinars discussing the future of Generative AI in enterprise software.
In 2023, Kiulian was named one of Forbes' 30 Under 30 in the Enterprise Technology category, recognizing his work with Colab and his contributions to the AI startup ecosystem. He has been quoted in industry publications on topics ranging from the ethics of Deep learning to the competitive landscape of AI infrastructure providers.
Current Work and Future Directions
As of 2024, Kiulian continues to serve as CEO of Colab, which has pivoted to focus on AI-powered project management tools for remote teams. The company has integrated support for multiple Large language model providers, including Anthropic's Claude and Google DeepMind's Gemini, allowing users to switch between models based on their needs. Colab reported annual recurring revenue of $1.8 million in the first quarter of 2024.
Kiulian has expressed interest in exploring Multi-Head Attention mechanisms for improving context retention in long-running conversations. He is also an advocate for open-source AI development, and Colab has released several of its internal tools for model evaluation under permissive licenses.
Personal Life
Kiulian is based in San Francisco, California. He is known to be an avid chess player and has participated in local tournaments. He has cited the strategic thinking required for chess as an influence on his approach to AI product design, drawing parallels between game-tree search and Beam Search decoding in language models.
He maintains an active presence on professional networking platforms, where he shares insights on AI entrepreneurship and occasionally mentors early-career engineers. Kiulian has stated in interviews that he is committed to ensuring that AI tools are accessible to non-technical users, a principle that guides his product decisions at Colab.