As one of the most promising emerging technologies, quantum computing has garnered increased attention from the research community. However, the quantity of available qubits and the short decoherence time severely restrict the capabilities of current quantum computers. We propose using a variety of approaches to tackle the challenge, including optimization for quantum optimum control, pulse-level compilation for variational quantum algorithms (VQAs), and a pulse-level abstraction layer for VQAs. The methods are superior than gate-level compilation methodology in terms of circuit latency and circuit fidelity. Pulse-level control offers us greater flexibility to optimize the compilation workflow. We can achieve comparable or better performance on VQA tasks while reducing latency by 70%.