Multimedia streaming, including on-demand and live delivery of content, has become the largest service, in terms of traffic volume, delivered over the Internet. The ever-increasing demand has led to remarkable advancements in multimedia delivery technology over the past three decades, facilitated by the concurrent pursuit of efficient and quality encoding of digital media. Today, the most prominent technology for online multimedia delivery is HTTP Adaptive Streaming (HAS), which utilises the stateless HTTP architecture - allowing for scalable streaming sessions that can be delivered to millions of viewers around the world using Content Delivery Networks. In HAS, the content is encoded at multiple encoding bitrates, and fragmented into segments of equal duration. The client simply fetches the consecutive segments from the server, at the desired encoding bitrate determined by an ABR algorithm which measures the network conditions and adjusts the bitrate accordingly. This method introduces new challenges to live streaming, where the content is generated in real-time, as it suffers from high end-to-end latency when compared to traditional broadcast methods due to the required buffering at client. This thesis aims to investigate low latency live adaptive streaming, focusing on the reduction of the end-to-end latency. We investigate the impact of latency on the performance of ABR algorithms in low latency scenarios by developing a simulation model and testing prominent on-demand adaptation solutions. Additionally, we conduct extensive subjective testing to further investigate the impact of bitrate changes on the perceived Quality of Experience (QoE) by users. Based on these investigations, we design an ABR algorithm suitable for low latency scenarios which can operate with a small client buffer. We evaluate the proposed low latency adaption solution against on-demand ABR algorithms and the state-of-the-art low latency ABR algorithms, under realistic network conditions using a variety of client and latency settings.