You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<p>InnerSource has been quietly shaping engineering culture for years -laying the groundwork for collaboration, reuse, and autonomy across teams. More recently, platform engineering has emerged with a similar mission, but from a different angle. Though their paths have long run in parallel, the connection between them has often been under-explored. In this session, we’ll reflect on how these two disciplines are beginning to converge, explore the shared principles that underpin their success, and consider how their reunion is reshaping the way teams collaborate, build, and share. This isn’t a deep dive into tools or implementation, but an invitation to rethink boundaries - and perhaps rediscover what’s been hiding in plain sight.</p>
<p>Generative AI isn’t just about creating—it’s about creating accurately, safely, and reliably. At IBM iX Events, Rahul architected the LLM Judge framework, a robust validation layer that evaluates AI-generated outputs autonomously, uses Theory of mind logic to learn from human in the loop to get better. He applied this in high-stakes environments like Wimbledon, the Masters, the Grammys, and ESPN NFL Fantasy. These systems handled 1,100+ RPS, served over 500,000 users live, and spanned 280+ ML models under intense, real-time loads. In this session, Rahul will walk through how LLM Judge was designed, scaled, and adapted across domains:</p>
<li>Domain adaptation: how we repurposed Wimbledon’s validation pipelines for golf commentary scoring, music summaries, and fantasy sport insights—all without starting from scratch.</li>
1664
+
<li>Reliability in action: infrastructure choices that sustained low latency and high throughput under peak event stress.</li>
1665
+
<li>Metrics that matter: including improving summary accuracy by 30%, reducing hallucination errors, and increasing user trust—as measured by engaged chat rates.</li>
Oscar has shared his insights at conferences such as DevOpsDays Medellín, DevSecOps Fest, and GitHub Copilot Week, and he was also featured in GitHub Latam Connect. His passion is enabling organizations to embrace InnerSource as a catalyst for innovation, compliance, and cultural transformation.</p>
<p><strong>Rahul Agarwal</strong> (Validating AI Outputs at Live Events)</p>
2030
+
<p>Rahul Agarwal is a Principal Data Scientist at IBM iX Events (New York), where he leads end-to-end generative AI systems for global sports and entertainment properties including Wimbledon, the US Open, the Masters, ESPN NFL Fantasy, and the Grammys. His real-time AI pipelines sustain 1,100+ requests per second, deliver content to 500k+ users, and draw on 280+ production ML models. An Emmy and IBM Corporate Technical Award winner, Rahul has authored 14+ patents and multiple publications on generative AI infrastructure and large-scale ML systems.</p>
<p>Robert is a passionate software developer and data engineer advocating for InnerSource at Bosch. He has joined the InnerSource initiative at Bosch in 2012. Robert is currently driving InnerSource adoption at Bosch as a member of the Center of Excellence Open and InnerSource, focusing on data driven insights and compliance.</p>
<p>Ryo has worked as a backend engineer and scrum master until 2018, and has been an engineering manager since. He is currently interested in InnerSource and promoting its adoption in his company.</p>
<p><strong>Sébastien Blanc</strong> (Platform Engineering and InnerSource: The Separated Twins Finally Reunited)</p>
2044
-
<p>Sébastien Blanc, Staff Developer Advocate at Port, is a Passion-Driven-Developer with one primary goal : share his passion by giving talks that are pragmatic, fun and focused on live coding.</p>
<p><strong>Shane Martin Coughlan</strong> (Understand Why Open Source Process Management Matters To InnerSource)</p>
2049
2055
<p>Shane Coughlan is an expert in communication, security and business development. His professional accomplishments include building the largest open source governance community in the world through the OpenChain Project, spearheading the licensing team that elevated Open Invention Network (OIN) into the largest patent non-aggression community in history and establishing the first global network for open source legal experts on behalf of Free Software Foundation Europe (FSFE). He is a co-founder of both the first law journal and the first law book dedicated to open source. He currently leads the OpenChain Project as General Manager and is a Staffing Committee, Management Board and General Assembly Member of OpenForum Europe.</p>
<p>InnerSource has been quietly shaping engineering culture for years -laying the groundwork for collaboration, reuse, and autonomy across teams. More recently, platform engineering has emerged with a similar mission, but from a different angle. Though their paths have long run in parallel, the connection between them has often been under-explored. In this session, we’ll reflect on how these two disciplines are beginning to converge, explore the shared principles that underpin their success, and consider how their reunion is reshaping the way teams collaborate, build, and share. This isn’t a deep dive into tools or implementation, but an invitation to rethink boundaries - and perhaps rediscover what’s been hiding in plain sight.</p>
<p>Generative AI isn’t just about creating—it’s about creating accurately, safely, and reliably. At IBM iX Events, Rahul architected the LLM Judge framework, a robust validation layer that evaluates AI-generated outputs autonomously, uses Theory of mind logic to learn from human in the loop to get better. He applied this in high-stakes environments like Wimbledon, the Masters, the Grammys, and ESPN NFL Fantasy. These systems handled 1,100+ RPS, served over 500,000 users live, and spanned 280+ ML models under intense, real-time loads. In this session, Rahul will walk through how LLM Judge was designed, scaled, and adapted across domains:</p>
<li>Domain adaptation: how we repurposed Wimbledon’s validation pipelines for golf commentary scoring, music summaries, and fantasy sport insights—all without starting from scratch.</li>
1653
+
<li>Reliability in action: infrastructure choices that sustained low latency and high throughput under peak event stress.</li>
1654
+
<li>Metrics that matter: including improving summary accuracy by 30%, reducing hallucination errors, and increasing user trust—as measured by engaged chat rates.</li>
Oscar has shared his insights at conferences such as DevOpsDays Medellín, DevSecOps Fest, and GitHub Copilot Week, and he was also featured in GitHub Latam Connect. His passion is enabling organizations to embrace InnerSource as a catalyst for innovation, compliance, and cultural transformation.</p>
<p><strong>Rahul Agarwal</strong> (Validating AI Outputs at Live Events)</p>
2019
+
<p>Rahul Agarwal is a Principal Data Scientist at IBM iX Events (New York), where he leads end-to-end generative AI systems for global sports and entertainment properties including Wimbledon, the US Open, the Masters, ESPN NFL Fantasy, and the Grammys. His real-time AI pipelines sustain 1,100+ requests per second, deliver content to 500k+ users, and draw on 280+ production ML models. An Emmy and IBM Corporate Technical Award winner, Rahul has authored 14+ patents and multiple publications on generative AI infrastructure and large-scale ML systems.</p>
<p>Robert is a passionate software developer and data engineer advocating for InnerSource at Bosch. He has joined the InnerSource initiative at Bosch in 2012. Robert is currently driving InnerSource adoption at Bosch as a member of the Center of Excellence Open and InnerSource, focusing on data driven insights and compliance.</p>
<p>Ryo has worked as a backend engineer and scrum master until 2018, and has been an engineering manager since. He is currently interested in InnerSource and promoting its adoption in his company.</p>
<p><strong>Sébastien Blanc</strong> (Platform Engineering and InnerSource: The Separated Twins Finally Reunited)</p>
2033
-
<p>Sébastien Blanc, Staff Developer Advocate at Port, is a Passion-Driven-Developer with one primary goal : share his passion by giving talks that are pragmatic, fun and focused on live coding.</p>
<p><strong>Shane Martin Coughlan</strong> (Understand Why Open Source Process Management Matters To InnerSource)</p>
2038
2044
<p>Shane Coughlan is an expert in communication, security and business development. His professional accomplishments include building the largest open source governance community in the world through the OpenChain Project, spearheading the licensing team that elevated Open Invention Network (OIN) into the largest patent non-aggression community in history and establishing the first global network for open source legal experts on behalf of Free Software Foundation Europe (FSFE). He is a co-founder of both the first law journal and the first law book dedicated to open source. He currently leads the OpenChain Project as General Manager and is a Staffing Committee, Management Board and General Assembly Member of OpenForum Europe.</p>
<p>InnerSource has been quietly shaping engineering culture for years -laying the groundwork for collaboration, reuse, and autonomy across teams. More recently, platform engineering has emerged with a similar mission, but from a different angle. Though their paths have long run in parallel, the connection between them has often been under-explored. In this session, we’ll reflect on how these two disciplines are beginning to converge, explore the shared principles that underpin their success, and consider how their reunion is reshaping the way teams collaborate, build, and share. This isn’t a deep dive into tools or implementation, but an invitation to rethink boundaries - and perhaps rediscover what’s been hiding in plain sight.</p>
<p>Generative AI isn’t just about creating—it’s about creating accurately, safely, and reliably. At IBM iX Events, Rahul architected the LLM Judge framework, a robust validation layer that evaluates AI-generated outputs autonomously, uses Theory of mind logic to learn from human in the loop to get better. He applied this in high-stakes environments like Wimbledon, the Masters, the Grammys, and ESPN NFL Fantasy. These systems handled 1,100+ RPS, served over 500,000 users live, and spanned 280+ ML models under intense, real-time loads. In this session, Rahul will walk through how LLM Judge was designed, scaled, and adapted across domains:</p>
<li>Domain adaptation: how we repurposed Wimbledon’s validation pipelines for golf commentary scoring, music summaries, and fantasy sport insights—all without starting from scratch.</li>
1653
+
<li>Reliability in action: infrastructure choices that sustained low latency and high throughput under peak event stress.</li>
1654
+
<li>Metrics that matter: including improving summary accuracy by 30%, reducing hallucination errors, and increasing user trust—as measured by engaged chat rates.</li>
Oscar has shared his insights at conferences such as DevOpsDays Medellín, DevSecOps Fest, and GitHub Copilot Week, and he was also featured in GitHub Latam Connect. His passion is enabling organizations to embrace InnerSource as a catalyst for innovation, compliance, and cultural transformation.</p>
<p><strong>Rahul Agarwal</strong> (Validating AI Outputs at Live Events)</p>
2019
+
<p>Rahul Agarwal is a Principal Data Scientist at IBM iX Events (New York), where he leads end-to-end generative AI systems for global sports and entertainment properties including Wimbledon, the US Open, the Masters, ESPN NFL Fantasy, and the Grammys. His real-time AI pipelines sustain 1,100+ requests per second, deliver content to 500k+ users, and draw on 280+ production ML models. An Emmy and IBM Corporate Technical Award winner, Rahul has authored 14+ patents and multiple publications on generative AI infrastructure and large-scale ML systems.</p>
<p>Robert is a passionate software developer and data engineer advocating for InnerSource at Bosch. He has joined the InnerSource initiative at Bosch in 2012. Robert is currently driving InnerSource adoption at Bosch as a member of the Center of Excellence Open and InnerSource, focusing on data driven insights and compliance.</p>
<p>Ryo has worked as a backend engineer and scrum master until 2018, and has been an engineering manager since. He is currently interested in InnerSource and promoting its adoption in his company.</p>
<p><strong>Sébastien Blanc</strong> (Platform Engineering and InnerSource: The Separated Twins Finally Reunited)</p>
2033
-
<p>Sébastien Blanc, Staff Developer Advocate at Port, is a Passion-Driven-Developer with one primary goal : share his passion by giving talks that are pragmatic, fun and focused on live coding.</p>
<p><strong>Shane Martin Coughlan</strong> (Understand Why Open Source Process Management Matters To InnerSource)</p>
2038
2044
<p>Shane Coughlan is an expert in communication, security and business development. His professional accomplishments include building the largest open source governance community in the world through the OpenChain Project, spearheading the licensing team that elevated Open Invention Network (OIN) into the largest patent non-aggression community in history and establishing the first global network for open source legal experts on behalf of Free Software Foundation Europe (FSFE). He is a co-founder of both the first law journal and the first law book dedicated to open source. He currently leads the OpenChain Project as General Manager and is a Staffing Committee, Management Board and General Assembly Member of OpenForum Europe.</p>
0 commit comments