This paper addresses the problem of assigning heterogeneous tasks with multiple skill requirements in crowdsourcing platforms. The aim is to find mutually exclusive, a highly-productive set of workers who can successfully complete the tasks within a given deadline and budget. We propose a timeline based weighted aggregation (TWA) technique to quantify the per-skill score of a worker. The score is computed based on the worker's profile and past work experiences. Given the worker's score, we formulate the problem as one of maximizing the productivity of all the N given tasks. A two-stage approximation solution is proposed. In the first stage, we offer a greedy-based 2-approximation algorithm for a single task. In the second stage, a local ratio based algorithm is proposed to extend the solution for multiple tasks. The overall solution is shown to be 3-approximate. Finally, simulation results using real-world data are presented to highlight the efficacy of our proposed schemes.