Repository navigation
Welcome to Agentic-workflows Discussions! #1
Replies: 15 comments 1 reply
|
Hi everyone! |
|
Hi all. I am co-chairing the agentic workflow working group alongside Ari. I have been interested in agents and agentic systems for a few years now. The AutoGPT paper/project caught my interest back in early 2023 and I have been tinkering/building these systems ever since: |
|
Hi everyone! My name is Marian Klose, I am a final year PhD Student in the PharMetrX program at Freie Universität Berlin, Germany. I have some limited experience in programmatic access to LLMs from python, and I am currently trying to get my feet wet with agentic LLM-driven workflows in Python using LangGraph. For the sake of my own learning, I sometimes post and plan to post irregularly on my little blog: https://marian-klose.com/posts/hitl_llm_agentic_pmx_workflow/. I am really interested in learning how we can use LLMs to speed up some clearly defined tasks in pharmacometrics without losing the critical validation process of each step that is needed in our domain. Looking forward to some interesting exchange in the future! |
|
Hi everyone, I’m Ari Pritchard-Bell. I’m a Scientific Associate Director at Amgen working at the intersection of agentic AI and pharmacometrics, and I’m co-chairing this group with Peter. Most of my time goes to Modus, an agentic orchestration framework for end-to-end pharmacometrics workflows. It’s one of the reference implementations we’ll curate here, but I’m just as interested in the layer underneath any single tool: the task definitions, benchmarks, and shared standards that outlast whatever harness runs them. I write about this at aripritchardbell.com, including a PharmAI series and some posts on Modus if you want background. Looking forward to meeting everyone. |
|
Hi all, my name is Se Jin Kim. I'm a pharmacometrician at Eli Lilly. Got my hands dirty with claude code and fell in love with it. I focus on constructing agentic workflows to ensure deterministic and reproducible outcomes. Hoping to gain wisdom from you all! |
|
Hi all, my name is Jason Sherfey. I'm a PMx/QSP modeler at BMS with a background in building and applying mechanistic modeling and AI/ML tools. I've been using agentic AI in two relevant ways that I differentiate: (1) modeling workflows for more-standard PK/PD and popPK analyses, (2) modeling workflows for less-standard/larger PBPK/QSP research and model development/application efforts. I think it will be interesting to see how lessons learned in the former domain can be extended/applied to the latter where applicable, especially as hybrid modeling approaches increase and accelerate with agentic AI (e.g., combining mechanistic PK/PD + NLME). Looking forward to working with you all! |
|
Hi everyone . I’m Brian Corrigan, a clinical pharmacologist and pharmacometrician currently serving as a Principal Consultant at Metrum Research Group. Prior to Metrum, I served as the Head of Translational Sciences at Pfizer, including the CP and PMX groups, Bionalytics, Imaging, Biomarkers and early clinical groups across TAs. I have been involved with ISoP for many years since its inception, including service as Board Member, President, Past-President, Chair of the ISoP 5-Year Strategic Plan Steering Committee, and ACoP Programming Chair. I was honored to be named an ISoP Fellow and to receive the ISoP Leadership Award. I have been playing around with developing agents via tools like Replit and Codex, and with integrating LLMS into MIDD-based shiny apps through API calls. For this SIG, I’m particularly interested in 1) helping to define and develop open-source skill files and workflows, 2) developing domain-specific benchmarks for evaluating LLM-based systems, and 3) helping other ISoP members see that AI/ML tools can build on and integrate with (not dilute or eliminate) the scientific basis of our disciplines, and help us bring more medicines to patients faster. |
|
Hi everyone! I am Elba Raimundez, I am a pharmacometrician at Sanofi based in Germany. My background is actually not pharmacometrics, but systems biology (intracellular signaling) and biochemistry/bioinformatics. I am serving a communication lead at the AI/ML SIG and led the coding assitants session in the "AI/ML in Pharmacometrics: Hands-On Workshop and Regulatory Panel" satellite meeting at PAGE 2026. I mostly use coding assistants (for me it is Github Copilot) to build workflows and get in-line improvements on my code. In the last year, I have started to further engineer my agentic prompts to exactly try to get reproducible, controlable and trustable outputs. I would love to spend more time in this, but I guess like many others in this group, sometimes it is difficult. That is why initiatives like this one are so important, so that we can all connect and learn from each other. That is why I am all-in 🤝 |
|
Hey everyone! My name is Sam Callisto, I am a PKPD Group Leader at MetrumRG where I've been for the past 7 years. During grad school I became interested in AI/ML and have continued to spend time researching its applications in PMx ever since. Despite being a relatively late adopter of LLMs, I have been contributing to the LLM tooling that we have been developing at MetrumRG for the past year and a half. This technology is already transforming our profession, so I want to participate in this WG to both learn from what others are doing and share the approaches that my team is exploring. |
|
Hey everyone, I'm Jean-Baptiste, AI Engineer at Nova In Silico, based in France. My work focuses on integrating AI into our in silico clinical trial simulation platform to help streamline clinical trials using disease models and digital patients. Over the last 2 years, I worked a lot on accelerating QSP model creation using human-assisted, informed AI agents, which is both exciting and challenging! Those days, I focus my work on writing skills for QSP modeling to better informed agents on using our QSP platform. Benchmark time is coming soon! You can already check our open source skills at https://github.com/novainsilico/jinko-skills . I'm really excited to share knowledge through this WG, exchange on good practices we want to adopt, and collectively explore where AI trends can generate real value in our day-to-day pharmacometrics work. Looking forward to exchanging more with you all! |
|
Hi Everyone, I'm Sai Ahil Palanisamy SD, PhD Scholar from the Centre for Pharmacometrics, Manipal, India. I'm currently working on the integration of Machine Learning approaches into quantitative pharmacology with main focus on covariate modeling. I have been amazed by the capabilities of the integration of AI into our work, especially the automation which now takes away the major part of work we usually did earlier. I'm really excited to join this group, learn and help in anyways possible! |
|
Hi everyone, My name is Shihao Hu, and I’ve been working as a Principal Scientist in Pharmacometrics at Johnson & Johnson for over three years. My day-to-day work focuses mainly on popPK/PD and exposure–response analyses. I’ve already seen significant efficiency gains from using LLM-based tools, and I’m particularly interested in how they can further automate some modeling and analysis workflow, allowing us to spend more time on creative (complex) research problems. Looking forward to learning from everyone and contributing to the group. |
|
Hi Everyone, |
|
Hi all, My name is Tim, and I am a MS graduate in Applied Mathematics (Dynamical Systems and Chaos) from SDSU. What brought me to PK/PD is my dry-lab research in "HIV Latent Infection in Morphine Exposure during Antiretroviral Thearpy (ART) treatment", that I spent doing for ~2.5 years that began from my undergrad and became my masters thesis; this work is heavily focused on QSP analysis from looking at ART optimizations of different treatment windows. I've been interested in extending my computational skillset into AI/ML from coding in MATLAB, Python, and now R. I genuinely enjoy doing the analysis of the computational results to understand deeply of the model or structure and interpreting the results into a translational aspect. |
|
Hi everyone! I’m JD Podichetty, Executive Director of AI and Predictive Analytics at the Critical Path Institute (C-Path). I’ve been exploring agentic workflows for MIDD, with a particular interest in making these approaches useful in day-to-day scientific work. I’m looking forward to learning from the group, exchanging ideas, and sharing both the successes and the “why did the agent do that?” moments. |
Uh oh!
There was an error while loading. Please reload this page.
👋 Welcome!
We’re using Discussions as a place to connect with other members of our community. We hope that you:
build together 💪.
To get started, comment below with an introduction of yourself and tell us about what you do with this community.
All reactions