We introduce a randomized Hall-Littlewood RSK algorithm and study its combinatorial and probabilistic properties. On the probabilistic side, a new model -the Hall-Littlewood RSK field -is introduced. Its various degenerations contain known objects (the stochastic six vertex model, the asymmetric simple exclusion process) as well as a variety of new ones. We provide formulas for a rich class of observables of these models, extending existing results about Macdonald processes. On the combinatorial side, we establish analogs of properties of the classical RSK algorithm: invertibility, symmetry, and a "bijectivization" of the skew-Cauchy identity. arXiv:1705.07169v1 [math.PR] 19 May 2017 1.4. Integrable degenerations. Section 4 deals with some available degenerations of the Hall-Littlewood RSK field.At the beginning, we identify the well-known objects -the stochastic six vertex model (introduced in [GS] and recently studied in [BCG]) and ASEP as degenerations of this field which appear when one considers the first column of random Young diagrams from the field only. This provides a new proof of the main result of [BBW] which relates first columns of Young diagrams distributed according to Hall-Littlewood processes and the height function of the stochastic six vertex model (see also [Bor] for connections between one point distributions of Macdonald measures and certain vertex models). As a corollary, results of Section 2 provide formulas for observables of these models. These formulas are essentially equivalent to the ones obtained in [TW1], [TW2] for ASEP with step initial condition, and in [BCG], [BP2] for the stochastic six vertex model by different methods, yet their derivation through Macdonald processes might be conceptually important.Full Young diagrams from the Hall-Littlewood RSK field contain more information than just their first columns, and thus provide more integrable models. In this paper we introduce the multi-layer stochastic vertex model and the multi-layer ASEP, which are natural generalizations of the stochastic six vertex model and ASEP, respectively. For any k ∈ Z ≥1 a k-layer stochastic vertex model appears when one restricts the Hall-Littlewood RSK field to the first k columns of Young diagrams. Results of Section 2 provide observables for these models. We mainly focus on the two-layer stochastic vertex model for which we give an explicit description (see Section 4.4).We use the term "multi-layer" for our models by an analogy with [PrSp] (see also [BO], [OR]) where similar multi-layer objects were introduced in the case of the classical RSK algorithm related to Schur functions. The multi-layer ASEP seems to be of a quite different nature than the extensively studied multi-species ASEP (see e.