Pune, the builders are in the room, the laptops are open, and the deep dives into agentic AI are about to get very real.
Today, DevSparks Pune 2026 will bring together developers, GCC leaders, product thinkers, and AI practitioners for a full day of deep technical exploration and honest dialogue around what's changing in software engineering.
The focus this year is sharp: AI in production. The day-long event offers a deep dive into real systems, real workflows, and real career implications, opening with a lightning talk that cuts straight to the heart of today's developer anxiety.
From writing to directing code: The new software engineer
Agentic AI and vibe coding are everywhere. But what actually changes for engineers? Which skills are quietly becoming obsolete? Which ones are compounding in value?
Arsh Goyal, AI & Engineering Influencer, will break down how developers can reposition themselves for a future where writing code is only part of the job, and directing systems becomes core leverage.
Indian GCCs shaping the global enterprise: What it means for tech teams
India's Global Capability Centres are no longer just support or innovation units. They are increasingly global decision nodes shaping enterprise tech strategy.
What does this rise in GCC importance mean for engineers on the ground? How are AI mandates flowing through global organisations? And how do Indian tech teams build influence, not just output?
Speakers include Amod Deshpande, Head of India, Engineering & Operations, Allvue Systems, and Anil Kasalanati, India Head of US Technology, MetLife.
Getting agentic apps ready for production: Lessons in observability and evaluation
Building a demo is easy. Operating an agentic system in production is not.
So, how do you trace decisions inside multi-step reasoning flows? How do you evaluate behavior, detect drift, and build feedback loops that scale?
Anannya Roy, Developer Advocate, Gen AI at Amazon Web Services, will walk through practical lessons in observability, evaluation frameworks, and production-grade reliability for agentic applications.
Building multi-agent workflows with Microsoft Agent Framework
Agent orchestration has moved from being just a research concept to a design pattern.
In this hands-on session, developers will explore open-source frameworks to design and deploy multi-agent systems across AI applications.
Mohammed Nawaz Shaikh, Cloud & AI Specialist - Digital Natives at Microsoft, will unpack how to move from experimentation to structured multi-agent architectures.
Accelerating GenAI with NVIDIA: From development to impact
Speed matters, especially when working with billion-parameter models. This session spotlights NVIDIA DGX Spark, a desk-side personal supercomputer that enables local prototyping over large models in hours, without cloud queues or runaway data costs.
Expect a technical walkthrough of how it integrates with CUDA, TensorRT, NeMo, RAPIDS, and NIM microservices for rapid agentic AI iteration. Sunil Patel, Manager, Solutions Architecture & Engineering, NVIDIA, will lead the deep dive.
Driving better outcomes for product teams with AI
How can spec-driven development and context engineering unlock better results for teams using AI daily?
Anand Hariharan, Co-founder & Chief Solutioning Officer at Indexnine Technologies, will share practical frameworks from real-world implementations.
Context in the age of AI: The new role of the engineer
As AI systems become more autonomous, clarity becomes currency. This means organisations should structure knowledge and teams before AI enters the workflow. Therefore, what does it mean for engineers to move from writing code to defining context?
Speakers include:
- Hiroshi Ito, CTO, Loglass
- Arjun Basu, VP of Engineering, Rakuten India
- Ryotaro Tsukii, Head of GCC Strategy & Service, Tech Japan Lab.
Agentic AI impact case study: Building sovereign healthcare resilience
The day closes with a real-world implementation story at population scale.
One of India's first sovereign healthcare foundation model integrates federated learning, knowledge graphs, and agentic AI to bridge decision support with innovation, securely and at scale.
Gaurav Tripathi, Co-founder and Group CTO at Partex.AI, will walk through the architecture, data design, and impact potential.
By the end of DevSparks Pune 2026, one thing will be clearer: builders who understand context, orchestration, and system thinking will define what comes next.

