Access to high-quality computer science education remains uneven across the United States, with rural schools facing persistent barriers including limited staffing, outdated infrastructure, and constrained professional development opportunities (Gutierrez & Terrones, 2023; Showalter et al., 2019). Computer Science in Rural California (branded as CS4NorCal), funded by a U.S. Department of Education's Education Innovation and Research (EIR) grant, sought to address these disparities by implementing a two-year professional development (PD) for teachers and school leaders to support the implementation of CS instructional pathways. The study targeted grades 4-8 and evaluates the impact of CS instruction on student achievement in mathematics and science and examines fidelity of implementation. Using a non-randomized cluster quasi-experimental design (QED) with a delayed treatment comparison group, the study analyzed outcomes from 29 schools. Treatment schools implemented a minimum of 30-50 hours of CS instruction. Average school achievement was measured using students' distance-from-standard (DFS) scores on California standardized assessments for mathematics and science. The study QED design meets What Works Clearinghouse (WWC) standards with reservations. Hierarchical linear modeling (HLM) to estimate program impacts. There was a statistically significant on science achievement (ES=0.24, p=0.01), with students in treatment schools narrowing the DFS gap by an average of five points compared to peers in comparison schools. The effect on mathematics achievement was positive but not statistically significant (ES=0.01, p=0.87). Although half of the measured indicators of fidelity within each leader and teacher component were met, overall fidelity of implementation was only met for the teacher professional development component in year 1. Findings support the study's premise that early integration of CS instruction within intermediate grades can be a viable strategy for improving academic outcomes in rural settings. Fidelity of implementation was assessed for the leadership and teacher professional development components. While the project achieved a relatively high level of fidelity with delivering planned program components, participant engagement varied, particularly among school leaders and counselors. Teacher and leaders were most engaged in the first year summer institute series, however robust engagement in ongoing, year-round planned professional learning declined in the second year. The decline in engagement reflected a number of challenges, including contextual and external constraints such as staffing shortages, extended responsibilities and limited time of rural educators, and COVID-19, as well as the poor alignment of content to rural classroom contexts. Despite inconsistent fidelity on participant indicators, collectively, findings suggest that the mechanisms of CS teacher learning and instructional opportunity were sufficiently activated to influence student engagement and learning outcomes. However, the study also highlights the limitations of turnkey professional learning models in rural contexts, where educators require flexible, context-sensitive support to adapt CS content to varied instructional formats. Future efforts can build on these findings to consider how we can design and promote systems of professional learning that can better support rural school leaders with broadly and systemically building instructional capacity in ways that support STEM-CS interdisciplinary teaching and learning across grades and subject matter.