🎓 Postgraduate Researcher @ IIIT Hyderabad
🧠 LLMs · VLMs · Multimodal AI · Indic NLP
🚀 Building Scalable AI Systems for Synthetic QA, Visual Reasoning, and Document Understanding
💻 Contributor to
OpenAI GPT-OSS •
Qwen 3 VL 235B •
Hugging Face Datasets
✨ I began my open-source journey with a merged PR to OpenAI’s GPT-OSS in May 2025, enabling vLLM Online Serve for OpenAI models.
Since then, I’ve been actively contributing to top-tier AI repositories under @hrithiksagar-tih, focusing on LLM/VLM Infrastructure, Multimodal AI, and Synthetic QA Generation.
Researcher · Engineer · Builder of Scalable AI Systems
🎓 MS by Research — Computer Science & Engineering
- 🏛️ IIIT Hyderabad
- 📅 2025 – Present
- 📧 Hrithik.Rachakonda@research.iiit.ac.in
- 👨🏫 Advisor: Prof. Ravi Kiran (@ravika)
🔬 Research Associate — BharatGen
- 📅 Dec 2024 – Present
- 📧 Hrithik.Sagar@tihiitb.org
- 🧭 Focus: Synthetic QA Generation, Indic Multilingual AI, Model Routing, Visual QA
- ⚙️ Building LLM/VLM Research Infrastructure
- 📅 May 2024 – Dec 2024
- 🧠 Bhashini Project (bhashini.gov.in)
- 🌐 Work: Indic NLP, Responsible AI, Cross-lingual QA
Spring 2024
- 🤖 Robotics: Planning and Navigation
- ⚖️ Responsible and Safe AI Systems
Fall 2023
- 🧠 Deep Learning: Theory and Practice
- 🖼️ Digital Image Processing
💼 Machine Learning Engineer (R&D) — March 2022 – Present
Research and development in LLM, VLM, vLLM, and AI Infrastructure
Specializing in Multi-GPU, Distributed Systems, and Data Pipelines
| Category | Tools |
|---|---|
| 💡 AI/ML | LLMs, VLMs, NLP, Deep Learning, Fuzzy Logic |
| ⚙️ Frameworks | PyTorch, TensorFlow, HuggingFace, vLLM, Ray |
| 🧠 Ops & Infra | Multi-Node GPU, Slurm, Ray, DuckDB, SQL |
| 🧩 Development | Python, VSCode, Git, Streamlit, Gradio, React |
| 🎨 Creative | Procreate (Digital Art), Shapr3D (3D Design) |
💡 Total: 4+ Merged PRs across top-tier AI repositories (OpenAI, Qwen, Hugging Face)
Contributed to the Offline Serve for vLLM feature
🔗 Merged PR #150
Improved Online & Offline Inference with vLLM
Enhanced dataset structure for Document Object Grounding & Reasoning (DOGR)
🔗 Merged PR / Discussion
| Repository | ⭐ Stars | 📦 Downloads | ✅ PRs Merged | 🚀 Contribution |
|---|---|---|---|---|
| OpenAI GPT-OSS | 18.3k | 8.9M/mo | 1 | Offline Serve for vLLM |
| Qwen 3 VL 235B | 13.2k | – | 2 | Online/Offline Inference |
| DOGR Dataset | – | – | 1 | Dataset Fixes |
