Speaking
Since 2020, I have spoken about AI, machine learning, observability, and engineering practice at conferences and organizations across India.
My strongest talks connect implementation detail to a larger operating question: how should teams build, evaluate, observe, and govern AI systems that people can depend on? Recordings of public sessions are linked below.
Events
Some of the events and organizations I’ve spoken at:
- ITER, Bhubaneswar: 2026
- PyConf Hyderabad: 2026
- AI Day Bhubaneswar: 2025
- TensorFlow User Group Bhubaneswar: 2025
- Odisha AI: 2020, 2022, 2024
- UCE Burla (faculty session): 2022
- Odisha Corporate Foundation: 2022
Bio for talks
If you need a short bio to introduce me, you’re welcome to modify this.
Soumendra Kumar Sahoo is an AI Observability Architect at PepsiCo with enterprise engineering experience since 2012. His career spans mainframes, Big Data, multi-cloud architecture, AIOps, enterprise search and MLOps/LLMOps.
His work focuses on the operating layer around production AI: platform strategy, observability, evaluation, reliability, and the engineering mechanisms that help teams deliver with confidence. He holds an M.Tech in Data Science and Engineering from BITS Pilani, is OpenTelemetry certified, and has published IEEE research on LLM fine-tuning.
Soumendra also maintains open-source AI and observability tooling, helps organize technical conferences, and contributes to public-knowledge communities.
Topics I speak about
- Building an operating model for production AI
- AI observability as a leadership and governance mechanism
- Evaluation, reliability, and responsible autonomy for AI agents
- MLOps and LLMOps platform strategy
- Leading technical communities without formal authority
- Open source and language technology for underrepresented communities
AI observability resources
My AI observability guide connects the talks to public walkthroughs, code and tracing examples.
| Session | Public resources |
|---|---|
| Developing an SDK for AI Agent’s Observability, PyConf Hyderabad, 15 March 2026 | Organizer session, written conference recap |
| Observability in AI, AI Day Bhubaneswar, 11 October 2025 | Interactive workshop, event page, written AI observability introduction |
Logistics
- Presentation: I mostly present interactive demos. Slides are secondary and mostly used for the introduction part.
- Laptop: I present from my own laptop (MacBook, HDMI). Please ensure a working HDMI connection at the podium.
- Aspect ratio: I can work with 16:9 or 16:10 screens. Please verify circles render as circles, not ovals.
- Internet: Some demos need an internet connection. A stable venue WiFi is preferred; I can fall back to mobile hotspot if needed.
- Recording: I encourage recording my talks. Please share the recording with me after the event.
- Attire: I dress casually. Please allow this.
Fee / honorarium
- I don’t charge a fee for public community talks. I accept Wikipedia donations, or GitHub Sponsors as honorarium. For corporate or private engagements, please contact me to discuss.
- I share recordings of my public talks openly.
Talk timeline
2026
- 2026-04-04: Tips & Tricks for using Claude Code at ITER, Bhubaneswar on the occassion of the annual techfest 2026 resumed after the interval of SIX years.
- 2026-03-15: Developing an SDK for AI Agent’s Observability at PyConf Hyderabad 2026.
2025
- 2025-10-11: Observability in AI at AI Day Bhubaneswar.
- 2025-06-08: Foundation of RAG to TensorFlow User Group, Bhubaneswar.
2024
- 2024-10-05: Model Monitoring in MLOps at Odisha AI conference.
2022
- 2022-11-05: Python Programming Language for Machine Learning to the faculties of UCE Burla.
- 2022-10-01: Demoed OdiaNLP, Mozilla Common Voice and OpenOdia open-source projects at Odisha AI conference.
- 2022-06-18: Machine Learning Tools & Libraries to Odisha Corporate Foundation.
2020
- 2020-12-04: Challenges in Machine Learning on a panel discussion at the first Odisha AI conference.