Entrepreneurship support organizations help startups survive and grow by compensating for market failure (e.g. inefficiencies in resource allocation), fostering strong business practices, and creating a fruitful environment for business exchange. These organizations arguably play a central role in innovation and business development and many of these initiatives are typically publicly, rather than privately funded. Typical representations of entrepreneurship support organizations are incubators, accelerators, technological transfer offices, and science parks. Recently, a new type of actor has emerged in the entrepreneurial ecosystem called venture builders. Venture builders are privately funded and repeatedly engage in new firm creation and support in collaboration with individual entrepreneurs. Research on the phenomenon of venture builders is still limited. This thesis therefore aims to add to our understanding of venture builders, and how venture builders may extend our understanding of entrepreneurship support organizations. Venture builders primarily work with startups - small and scalable businesses with high growth potential, often within the technology sector. In recent years they have gained significant interest among practitioners. Successful unicorns such as Zalando, Moderna, Snowflake, and Hello Fresh, have emerged from venture builders’ activities and engagement. This research aims to enhance our understanding of venture builders and their role as entrepreneurship support organizations. Theoretically, this research contributes to the entrepreneurship support literature by shedding light on venture builders as for-profit actors having a strategic intent to engage in new venture creation. This is done by following structured methodologies, controlling and committing resources, as well as orchestrating internal and external networks, aiming at fostering successful entrepreneurship. This thesis argues that venture builders may be referred to as a distinct type of entrepreneurship support organization.

In the rapidly evolving landscape of artificial intelligence (AI) and machine learning (ML), transitioning from theoretical models to robust, adaptive production systems remains a significant challenge in the digital age. As data-driven methodologies revolutionize problem solving across industries, a critical gap exists between research achievements and reliable real-world deployment. Key obstacles include concept drift and data quality issues that arise from unpredictable changes in data distributions and operational complexities that systematically affect model performance, reliability, and efficiency. This thesis addresses these challenges by introducing an overarching framework for adaptive ML systems that operate reliably in dynamic, real-world environments, incorporating innovative methodologies for dynamic drift detection and real-time data quality assessment bounded by robust Machine Learning Operations (MLOps) strategies. These integrated components enable the creation of production-grade ML systems that can efficiently adapt to shifts in data distributions and assess data quality in real-time, ensuring stable and reliable performance in dynamic environments. The proposed approaches are validated through real-world use cases, demonstrating significant improvements in predictive accuracy and operational efficiency. By deploying these adaptive systems in industrial contexts, the thesis highlights their potential to deliver reliable, high-performance ML solutions tailored to the demands of complex time-sensitive applications. This work offers concrete solutions for translating theoretical advances into practical applications, contributing to developing robust and scalable ML systems for real-world deployment.

Exhaustion disorder (ED) is a relatively new diagnosis associated with cognitive symptoms, which are normally assessed using standardized cognitive tests or questionnaires targeting everyday cognitive failures (i.e., subjective cognitive complaints, SCCs). The purpose of this thesis was to add empirical knowledge on the self-reported, first-hand experience of cognitive function in ED. More specifically, it aimed to learn how SCCs relate to test performance, psychological distress and neural activity. Further objectives were to evaluate the types of difficulties being expressed, and what aspects can be considered helpful or hindering with respect to cognitive recovery. 

Study 1 found that when compared to healthy controls, ED patients reported substantially higher levels of SCCs, and were more likely to express difficulties in situations without external memory cues. In both groups, the level of SCCs was correlated with psychological distress and not with cognitive test-results. Using functional magnetic resonance imaging, Study 2 investigated the relationship between SCCs, test performance and brain activity. There was no association between SCCs and behavioural results on the in-scanner task, tapping response inhibition. However, a positive correlation was detected between SCCs and relatively more brain activity in a cluster in the right-side occipital lobe during the more difficult task condition. This exploratory finding may indicate compensational neural activity, possibly involving visual processing or the altering between task positive and task negative neural networks. Study 3 analysed interviews with people who had participated in ED-rehabilitation 6-10 years earlier, and displayed a range of individual experiences. Cognitive symptoms had been highly distressing. Lingering problems were also noted in several cognitive areas, but maintenance of attention and executive control may be  domain-general areas of importance. Cognitive recovery was seen as closely tied to context, including the overall life situation and general recovery from ED, which varied between individuals. Hence, different restorative or compensatory strategies were considered helpful, as were optimization of the external environment and a change in approach towards the own self and cognitive performance.

In sum, this thesis studied the subjectively reported cognitive symptoms in ED. It has supported and extended previous findings by showing how substantial cognitive difficulties may be experienced, and that the expression of these problems is intricately linked to different facets and levels of cognition.

The rapid rise of Cellular Internet of Things (CIoT) technology is expected to connect over 6 billion devices by 2029. Many of these devices, often deployed in remote, urban, or hard-to-reach areas, operate on limited battery resources and are expected to last up to 10 years. However, current battery limitations challenge the long-term operation required by many applications. Ensuring low energy consumption is therefore crucial for avoiding frequent recharging or battery replacements.

This thesis addresses the challenge of enhancing the energy efficiency of Narrow-Band Internet of Things (NB-IoT) devices by examining and optimizing the energy-saving mechanisms standardized by the 3rd Generation Partnership Project (3GPP). Specifically, the research classifies and evaluates existing energy-saving solutions for CIoT— particularly for NB-IoT—by identifying their limitations and studying the impact of mechanisms such as Discontinuous Reception (DRX), Release Assistance Indicator (RAI), Power Saving Mode (PSM), Early Data Transmission (EDT), and Preconfigured Uplink Resources (PUR) on battery life. While improved energy efficiency is essential, it often comes at the cost of increased latency. This thesis evaluates these effects on both energy consumption and latency, offering insights into the trade-offs between the two.

Based on these findings, we propose guidelines for configuring NB-IoT devices to achieve an optimal balance between energy efficiency and performance. A significant contribution of this research is the development of a machine learning-based optimization approach that dynamically adjusts configurations based on network conditions, such as signal quality, packet loss, and data transmission frequency. By integrating advanced energy-saving mechanisms with optimization techniques, this work deepens our understanding of the interplay between device configurations and battery life. Although energy-saving measures may reduce performance (e.g., increased latency or reduced throughput), further investigation into these trade-offs is essential. The proposed guidelines and strategies aim to extend NB-IoT devices’ battery life to 10 years or more, enhancing their usability across diverse CIoT deployments.