SRM University, AP (11 May 2026- 13 May 2026) -
We are thrilled to announce the successful completion of yet another impactful Faculty Development Program (FDP), at the prestigious SRM University, AP.
This intensive 3-day program focused on Advanced AI & LLM Systems, diving deep into the technologies shaping the future of innovation. It was wonderful to engage with the brilliant faculty members and researchers at SRM University, exploring the practical applications and research potential of Large Language Models and Agentic AI.
At Global Infoventures, we are committed to bridging the gap between industry advancements and academic excellence. A huge thank you to all participants for their incredible energy and looking forward to many more impactful learning initiatives ahead!
Together, we are empowering the next generation of educators and innovators.
DAY 1
Setting the Stage for Innovation!
Day 1 of our Faculty Development Program (FDP) at 𝗦𝗥𝗠 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆, 𝗔𝗣 was a deep dive into the core of next-gen intelligence. We kicked off the series with a focus on 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀 𝗼𝗳 𝗠𝗼𝗱𝗲𝗿𝗻 𝗩𝗶𝘀𝗶𝗼𝗻 𝗔𝗜 𝗮𝗻𝗱 𝗦𝗰𝗮𝗹𝗮𝗯𝗹𝗲 𝗠𝗼𝗱𝗲𝗹 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴.
It was an incredible experience working with the faculty and researchers as we navigated the complexities of high-performance AI. Here’s a snapshot of what we covered :
* 𝗘𝘃𝗼𝗹𝘂𝘁𝗶𝗼𝗻 𝗼𝗳 𝗩𝗶𝘀𝗶𝗼𝗻 𝗔𝗜
We explored the journey from real-time detection architectures to 𝗮𝗱𝘃𝗮𝗻𝗰𝗲𝗱 𝗺𝘂𝗹𝘁𝗶𝗺𝗼𝗱𝗮𝗹 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀. The focus remained on building systems with enhanced contextual understanding and superior decision-making capabilities.
* 𝗛𝗶𝗴𝗵-𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 & 𝗢𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻
𝘛𝘩𝘦𝘰𝘳𝘺 𝘮𝘦𝘵 𝘱𝘳𝘢𝘤𝘵𝘪𝘤𝘦 𝘢𝘴 𝘱𝘢𝘳𝘵𝘪𝘤𝘪𝘱𝘢𝘯𝘵𝘴 -
- Performed 𝗰𝘂𝘀𝘁𝗼𝗺 𝗺𝗼𝗱𝗲𝗹 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 on domain-specific datasets.
- Leveraged 𝗵𝗶𝗴𝗵-𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗚𝗣𝗨 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 and accelerated workflows.
- Implemented optimization strategies to improve 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 and minimize latency for production-scale deployment.
* 𝗦𝗰𝗮𝗹𝗶𝗻𝗴 𝘃𝗶𝗮 𝗗𝗶𝘀𝘁𝗿𝗶𝗯𝘂𝘁𝗲𝗱 𝗦𝘆𝘀𝘁𝗲𝗺𝘀
To tackle the bottlenecks of large-scale AI, we delved into distributed GPU environments. From overcoming memory limitations to implementing data-level and model-level parallelism, we pre-trained generative transformer architectures to understand the blueprint of scalable AI system design.
The energy and technical curiosity from the participants at SRM University, AP was very encouraging for our team of trainers!
Stay tuned as we update on how we continued to push the boundaries of AI & LLM systems over the next two days.
DAY 2
𝗖𝗼𝗻𝘁𝗶𝗻𝘂𝗶𝗻𝗴 𝘁𝗵𝗲 𝗺𝗼𝗺𝗲𝗻𝘁𝘂𝗺 𝗮𝘁 𝗦𝗥𝗠 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆, 𝗔𝗣!of our Faculty Development Program (FDP) was all about speed, efficiency and the practical side of massive models. We shifted our focus to Accelerated Transformer Workflows and Efficient Adaptation of LLMs.
As the demand for Generative AI grows, the ability to adapt these models sustainably is becoming a critical skill for both researchers and industry leaders. Here’s what we explored -
* 𝗔𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗲𝗱 𝗧𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗿 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲𝘀
We moved beyond basic training to GPU-accelerated pipelines. By focusing on high-throughput workloads, we demonstrated how to achieve faster experimentation, training and inference — essential for modern deep learning tasks where time-to-market and compute costs are key.
* 𝗣𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿-𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗙𝗶𝗻𝗲-𝗧𝘂𝗻𝗶𝗻𝗴 (𝗣𝗘𝗙𝗧)
One of the highlights was our deep dive into lightweight fine-tuning strategies. Participants gained hands-on exposure to -
- Low-bit optimization and quantization.
- Instruction alignment to sharpen model accuracy.
- Domain adaptation techniques that maintain high performance while drastically reducing memory and computational requirements.
* 𝗦𝗰𝗮𝗹𝗮𝗯𝗹𝗲 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗠𝗼𝗱𝗲𝗹𝘀
We addressed the "real-world" challenge : How do you deploy these models on limited hardware? From quantized training strategies to efficient checkpoint handling, we explored how to customize large generative models for specialized enterprise and research applications without needing a supercomputing cluster at every turn.
It’s inspiring to see the faculty and researchers at SRM University, AP engaging with these complex workflows to build deployment-ready AI systems.
𝗗𝗮𝘆 𝟯
𝗮𝘁 𝗦𝗥𝗠 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆, 𝗔𝗣’𝘀 𝗙𝗮𝗰𝘂𝗹𝘁𝘆 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 𝗣𝗿𝗼𝗴𝗿𝗮𝗺 (𝗙𝗗𝗣) : 𝗗𝗲𝗹𝘃𝗶𝗻𝗴 𝗶𝗻𝘁𝗼 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗞𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗦𝘆𝘀𝘁𝗲𝗺𝘀 𝗮𝗻𝗱 𝗔𝘂𝘁𝗼𝗻𝗼𝗺𝗼𝘂𝘀 𝗔𝗜 𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀!
The final session marked a massive shift from standard conversational AI to the future of enterprise automation — moving from standalone models to autonomous compound AI systems.
Here are the key takeaways and hands-on milestones from Day 3:
* 𝟭. 𝗜𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝘁 𝗥𝗲𝘁𝗿𝗶𝗲𝘃𝗮𝗹-𝗗𝗿𝗶𝘃𝗲𝗻 𝗔𝗜 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲𝘀 Implemented context-aware pipelines enabling LLMs to interact seamlessly with external knowledge sources for more reliable, grounded and domain-aware responses.
Gained practical exposure to semantic retrieval workflows, vector-based knowledge indexing and real-time information integration.
* 𝟮. 𝗧𝗵𝗲 𝗥𝗶𝘀𝗲 𝗼𝗳 𝗔𝘂𝘁𝗼𝗻𝗼𝗺𝗼𝘂𝘀 𝗖𝗼𝗺𝗽𝗼𝘂𝗻𝗱 𝗔𝗜 Explored how AI is transitioning into systems capable of complex reasoning, orchestration, planning and multi-step task execution.
Studied modular AI workflow design featuring interconnected intelligent components, dynamic execution pipelines and adaptive decision-making.
* 𝟯. 𝗛𝗮𝗻𝗱𝘀-𝗢𝗻 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 Built and deployed 𝗮𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲𝘀 using both memory-enabled and stateless execution flows.
Integrated prebuilt and custom tools to push model capabilities far beyond simple text generation — turning them into engines for intelligent task automation.
* 𝟰. 𝗢𝗯𝘀𝗲𝗿𝘃𝗮𝗯𝗶𝗹𝗶𝘁𝘆 & 𝗥𝗲𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 Deep-dived into execution tracing, debugging and real-time monitoring techniques to evaluate complex AI interactions and ensure production-grade reliability.
Wrapped up the FDP with incredible energy as we bridged the gap between theoretical AI and robust, autonomous workflows ready for real-world deployment!
After an extremely rewarding expeerience over the 3 days FDP, we look forward to many such interactions with the incredible team at SRM University, AP.
