Future scope of the dataweave codegen project
Posted: Wed May 28, 2025 4:58 am
for the fine-tuned models, we see that an immediate scope for future work is to improve the pass@k for the codes that compile. This may be achieved by adding more high quality data (like the matched set that was used).
Additionally, we are investing in reinforcement learning methods like rlhf that provide signals to llm based on whether the generated code compiles and produces the right output.
Based on the observed trajectory, we believe we will soon outperform private third-party models for dataweave codegen by fine-tuning the salesforce xgen model. By doing so, we achieve the best pass@k metrics compared to private models such as gpt3.5/4 or claude, and the ongoing inference latency will be much lower given the smaller size of the xgen model.
This content was created in collaboration between the author and the america phone number list following individuals: yazdan jamshidi and hadi minooei.go behind the scenes with “ask the trailblazer” host jaylena lomenech to discover the magic formula of sharing knowledge to grow careers in the salesforce ecosystem.
As someone who made a career pivot into tech, I understand the impact that representation plays in fostering a growth mindset. At the beginning of my career, i leveraged trailhead—salesforce’s free online learning platform—my global network, countless tutorials, and workshops to learn about salesforce and how to have an impact in the technology sector.
Representation matters
but the biggest influence for me has come from the black leaders who inspire, motivate, and champion others through their work. The ones who make an impact through sharing their own experiences of being a person of color in tech.
Additionally, we are investing in reinforcement learning methods like rlhf that provide signals to llm based on whether the generated code compiles and produces the right output.
Based on the observed trajectory, we believe we will soon outperform private third-party models for dataweave codegen by fine-tuning the salesforce xgen model. By doing so, we achieve the best pass@k metrics compared to private models such as gpt3.5/4 or claude, and the ongoing inference latency will be much lower given the smaller size of the xgen model.
This content was created in collaboration between the author and the america phone number list following individuals: yazdan jamshidi and hadi minooei.go behind the scenes with “ask the trailblazer” host jaylena lomenech to discover the magic formula of sharing knowledge to grow careers in the salesforce ecosystem.
As someone who made a career pivot into tech, I understand the impact that representation plays in fostering a growth mindset. At the beginning of my career, i leveraged trailhead—salesforce’s free online learning platform—my global network, countless tutorials, and workshops to learn about salesforce and how to have an impact in the technology sector.
Representation matters
but the biggest influence for me has come from the black leaders who inspire, motivate, and champion others through their work. The ones who make an impact through sharing their own experiences of being a person of color in tech.