> Markdown version of [/@harshita-seth](https://www.wearedevelopers.com/@harshita-seth). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Harshita Seth
Solution Architect at Nvidia specializing in LLM compression and domain adaptation.

## About

Harshita Seth makes large language models run faster and cost less. As a Solution Architect at Nvidia, she focuses on model compression techniques—specifically quantization, pruning, and knowledge distillation.

Beyond shrinking model sizes, she helps engineering teams inject domain-specific knowledge and cultural nuances into open-source LLMs. Her work shows developers what to do when standard retrieval-augmented generation falls short for specialized business applications.



## Past Sessions

### World Congress 2025 · July 9, 2025
Berlin, Germany
- [NVIDIA Expert Session: Efficient Generative AI Inference Using Model Compression Techniques: Quantiz](https://www.wearedevelopers.com/events/world-congress-2025/sessions/575-nvidia-expert) · 60 min
- [Adding knowledge to open-source LLMs](https://www.wearedevelopers.com/videos/1522-adding-knowledge-to-open-source-llms) · 30 min · 🎥 Watch recording
- [ Model Compression Techniques for Efficient LLM Deployment](https://www.wearedevelopers.com/events/world-congress-2025/sessions/811-model-compression) · 120 min



## Videos

- [Adding knowledge to open-source LLMs](https://www.wearedevelopers.com/videos/1522-adding-knowledge-to-open-source-llms) · Harshita Seth, Sergio Perez