Background/Context: Despite increasing school enrollment, significant literacy gaps persist in developing countries, with many children unable to read by age 10. This paper evaluates "M'ap Li Net Ale," a comprehensive intervention aimed at improving early-grade reading skills in Haiti, a context characterized by limited government capacity and challenging socio-economic conditions. Purpose/Objective/Research Question: The study aims to assess the effectiveness of a bundled intervention, with a focus on structured pedagogy. The intervention included teacher and principal training, the provision of pedagogical material, and one-on-one teacher coaching, with the goal of improving reading proficiency among early-grade students. Setting: The intervention was implemented in a nationally representative sample of 446 public schools in Haiti over two academic years. Population/Participants/Subjects: The study follows first-grade students enrolled in the selected schools during the first year of the intervention, with an initial sample of approximately 6,000 students. During the second year, a new similarly sized cohort of first graders was added to the study, and those in treatment schools were treated. Intervention/Program/Practice: The "M'ap Li Net Ale" program aimed to improve early-grade reading skills through a multifaceted approach. It provided comprehensive training for teachers and principals, supplied classrooms with textbooks and structured lesson plans, and offered regular coaching sessions. Additionally, the program included community mobilization efforts and offered extracurricular activities for students. Recognizing the challenges of financial constraints and the impact of widespread insecurity on mobility, the study examined two distinct versions of the program to assess their scalability in such an environment: the "Full" version, which includes external coaching, and a lower cost "Light" version, where coaching is led by school principals. Research Design: We used a randomized controlled trial (RCT) to assess the intervention's impact. Schools were randomly assigned to either the "Full" intervention, "Light" intervention, or control group, allowing us to compare outcomes across different models. Data Collection and Analysis: The study used the Early Grade Reading Assessment (EGRA) to measure reading skills before and after the intervention, along with surveys to gather data on socio-economic conditions, school infrastructure, and implementation fidelity. Findings/Results: Eighteen months post-intervention, we found significant improvements in reading skills in both the "Full" and "Light" groups, with no statistically significant difference between the two, suggesting that the principal-led "Light" version was as effective as the more resource-intensive "Full" version. This suggests potential for scaling the intervention in settings with limited funding and low state capacity. Conclusions: The intervention had large and positive effects on student learning outcomes, with both the "Full" and "Light" versions increasing overall Early Grade Reading Assessment (EGRA) scores by 25 to 27 percent. Specifically, for the cohort that benefited from the intervention for two consecutive years, treated students in the Light and Full programs scored 7.2 (26.8 percent) and 6.8 percentage points (25.4 percent) higher, respectively, than the control group in our measure of performance. For the cohort that received the intervention for only one year, treated students scored 2.4 (16 percent) and 1.9 percentage points (13 percent) higher than the control group. These results indicate that the less costly "Light" approach, which relies on school principals for coaching, was as effective as the "Full" approach across all dimensions of learning assessed by EGRA. The study finds that substantial learning gains can be achieved through interventions operationalized within existing education systems, even in settings of limited capacity and challenging socioeconomic contexts.