PhD Researcher · Computer Science
Sree Bhargavi Balija
University of Missouri, Columbia · MS UC San Diego · BE IIT Hyderabad
I build and study federated learning systems, privacy-preserving ML, and agentic AI infrastructure — spanning LLM efficiency & quantization, differential privacy, decentralized multi-agent frameworks, neuromorphic computing, and trustworthy AI. My research asks how intelligent systems can collaborate without centralizing data, and how LLMs can be made efficient and trustworthy at scale. I publish at AAAI, NeurIPS, and collaborate with MIT Media Lab, Meta Research, and Stanford.
75+Citations
11Publications
3Fellowships
5+Collabs
01
News & Honors
🏆 MIT NANDA Radius Fellow 2025·
🎤 AAAI 2026 Oral Presentation·
📄 NeurIPS 2025 WIML Workshop·
🎤 AIM-2026 Oral Presentation·
🧠 EleutherAI Open AI Research Fellow·
📜 AAAI 2024 · AAAI 2025 · Stanford·
💡 NSF AI-SDM Workshop · CMU·
🔬 Algoverse AI Research Fellow 2025·
🏅 IMMO Silver Medalist·
🎓 Academic Excellence · IIT Hyderabad·
🏆 MIT NANDA Radius Fellow 2025·
🎤 AAAI 2026 Oral Presentation·
📄 NeurIPS 2025 WIML Workshop·
🎤 AIM-2026 Oral Presentation·
🧠 EleutherAI Open AI Research Fellow·
📜 AAAI 2024 · AAAI 2025 · Stanford·
💡 NSF AI-SDM Workshop · CMU·
🔬 Algoverse AI Research Fellow 2025·
🏅 IMMO Silver Medalist·
🎓 Academic Excellence · IIT Hyderabad·
2026
Oral
Oral presentation accepted at AAAI 2026 — presenting latest research on federated learning and agentic AI systems.
Sep 2025
Paper
Poster accepted at NeurIPS 2025 (WIML Workshop) — "Hybrid SNN-Transformer Networks for Event-Based, Energy-Efficient Large-Scale Learning."
Aug 2025
Oral
Oral presentation at AIM-2026, HPC session — "EcoFlow: An AI-Optimized GPU Orchestrator for Exascale Sustainability."
Aug 2025
Position
Joined Fujitsu Research of America as Software Engineering Intern. Working on GPU utilization and agentic AI integration in the Kojuchi platform.
Jul 2025
Fellowship
Selected as MIT NANDA Radius Fellow 2025 — researching decentralized agent discovery, trust, and microtransaction-based interoperability.
Jul 2025
Paper
Co-authored paper with MIT Media Lab, Meta, and Stanford — "An Index-based Federated Registry Architecture for Secure, Capability-Aware Agent Discovery."
Jun 2025
Workshop
Project accepted to NSF AI-SDM Workshop 2025 on Human-AI Complementarity for Decision Making (CMU).
May 2025
Fellowship
Full scholarship recipient — Algoverse AI Research Program 2025. Selected among top candidates.
Jan 2026
Workshop
Represented UC San Diego at the IEEE Rising Stars 2026 Project Showcase with "FedNAM+: An Interpretable and Uncertainty-Aware Federated Learning Framework."
May 2025
Collab
Contributed to Terminal Bench with Prof. Ludwig at Stanford — benchmarking tools for AI terminal applications.
Jan 2026
PhD
Started PhD in Computer Science at University of Missouri, Columbia.
Apr 2024
Paper
Published at AAAI 2024 (Stanford) — "Building Communication Efficient Asynchronous Peer-to-Peer Federated LLMs with Blockchain." SB Balija, A Nanda, D Sahoo. Proceedings AAAI Symposium Series 3(1), 288–292.
02
Publications
AILUMINATE: Introducing v1.0 of the AI Risk and Reliability Benchmark from MLCommons
Explore-Then-Commit with Dynamic Expertise Amplification
Hybrid SNN-Transformer Networks for Event-Based, Energy-Efficient Large-Scale Learning
Poster
Building Communication Efficient Asynchronous Peer-to-Peer Federated LLMs with Blockchain
The Trust Fabric: Decentralized Interoperability and Economic Coordination for the Agentic Web
Elastic MIG Reconfiguration with PCIe-Aware Placement for Multi-Tenant GPUs
03
Research & Industry Experience
Software Engineering Intern
Fujitsu Research of America · USA
- Optimizing GPU utilization in the ACB project via workload balancing and scheduling strategies.
- Implementing multi-server configurations to enhance parallel processing efficiency.
- Exploring agentic AI integration into Kojuchi platform for autonomous decision-making and workflow orchestration.
ML Researcher — Fellowship
MIT NANDA + Radius Collaboration · USA
- Designed Radius Testnet trust and discovery layer for agent marketplaces with smart contract–based reputation mechanisms.
- Implemented MCP server monetization workflows — payments, invoicing, EVMAuth, TypeScript SDK adapters.
- Researched agentic AI ecosystems: decentralized discovery, trust, and microtransaction-based interoperability.
Research Collaborator — Differential Privacy
Prof. Praneeth Vepakomma · MIT · Remote
- Built patient-specific differential privacy pipeline with DP-Prompt paraphrasing and privacy-preserving embeddings.
- Optimized embeddings per patient for scalable privacy analysis across multimodal clinical datasets.
Lead Researcher — Transformer Optimization
Meta Research Collaborator (CHAI Model) · USA
- Pioneered hybrid sparse attention + precision fine-tuning across 3 domain-specific datasets — +15% task accuracy, −50% inference time.
- Engineered memory-optimized transformer pipelines reducing GPU memory by 4GB+ per training cycle.
- Innovated AI-driven clustering for token embeddings, reducing pre-processing overhead by 40%.
Machine Learning Engineer
Akdene Technologies · USA
- AI-powered financial risk engine (TensorFlow, XGBoost, PyTorch Geometric) — real-time fraud detection for 100K+ customers.
- Agentic chatbot (GPT-4o, RAG, AutoGen on Django) — reduced support workload 120+ hrs/month, response time by 65%.
- Serverless AWS + Kubernetes deployment achieving 99.9% uptime; reduced API costs 30% via LoRA fine-tuning.
Software Engineer
ServiceNow · India
- Scalable backend microservices (Java, Spring Boot, REST) — 3× data processing improvement for enterprise clients.
- OAuth2 authentication system — reduced unauthorized access incidents by 60%.
- GCP cloud migration — reduced infrastructure costs 25%.
04
Selected Projects
Featured · MIT AI Summit 2025
AutoPatch+
Next-gen AI code validation platform with hallucination detection for generative coding. Co-founded and showcased at MIT AI Summit 2025.
LLMCode ValidationHallucination Detection
AAAI 2024
P2P Federated LLMs
Communication-efficient asynchronous federated learning for LLMs using blockchain-based decentralized coordination.
Federated LearningBlockchainLLMs
arXiv 2024
CPTQuant
Mixed-precision quantization for efficient LLM deployment on resource-constrained systems. Novel calibration techniques for post-training quantization.
QuantizationLLM EfficiencyMixed Precision
Hugging Face
PixPrompt
Multimodal small LLM (125M parameters) for efficient image-text generation. Designed for edge deployment.
MultimodalSmall LLMImage-Text
05
Education
Jan 2026 –
Present
Present
PhD in Computer Science
University of Missouri, Columbia
Federated Learning · Privacy-Preserving ML · Agentic AI Systems · LLM Efficiency
Sep 2022 –
Jun 2024
Jun 2024
MS in Machine Learning and Data Science
University of California, San Diego
Deep Generative Models · AI Learning Algorithms · Statistical Learning · Recommender Systems · Introduction to Deep Learning
Jul 2016 –
May 2020
May 2020
Bachelor of Engineering
Indian Institute of Technology, Hyderabad
Algorithms · Data Structures · Probabilistic Graphical Models · Artificial Intelligence · DBMS
06
Technical Skills
Languages
Python (Expert) · Java · JavaScript · C++ · R · Fortran · Kotlin · Swift
ML / Research
Federated Learning · NLP · Deep Learning · Multimodal AI · Conformal Predictions · Model Interpretability
Frameworks
PyTorch · TensorFlow · Hugging Face · LangChain · XGBoost · Scikit-learn · Django · Keras
LLM / Agentic
GPT-4o · LLaMA-2 · RAG · AutoGen · Multi-agent · NLTK · Google Gemini · OpenAI API
MLOps / Infra
Kubernetes · Docker · CI/CD · MLflow · Evidently · Feast · GitHub Actions · ArgoCD · Prometheus
Cloud / Data
AWS · GCP · Azure · FastAPI · gRPC · Elasticsearch · Kafka · Cassandra · Redis · SQL