Gayathri Akkinapalli

Hello! I'm Gayathri

I'm an AI/ML Engineer who loves turning ambitious ideas into systems that actually work. With a Master's in Computer Science from UMass Amherst (4.0 GPA) and four years of experience across research and industry, I've worked across the modern AI stack, from search and retrieval systems to LLM fine-tuning, RAG pipelines, and healthcare intelligence.

My work covers hybrid retrieval combining dense and sparse search with neural re-ranking, agentic frameworks for LLM personalization, RAG architectures, healthcare AI serving millions of users, and generative models for human motion synthesis. The motion synthesis research was published at IEEE WACV '25, and along the way I've also worked on LLM alignment, safety, and machine translation.

When I'm not designing multi-agent pipelines or fine-tuning models, you'll find me exploring new research papers and the latest in generative AI, mentoring fellow engineers, or occasionally setting Guinness World Records (yes, really—for Kuchipudi dance!).

"Always learning, always building—driven by the belief that AI should work for everyone."

Domains of Interest

Large Language Models

Fine-tuning, alignment, safety, and personalization using RLHF, DPO, and LoRA techniques.

RAG & Agentic Systems

Multi-agent architectures, knowledge retrieval, and autonomous AI pipelines.

Reinforcement Learning

Policy optimization, actor-critic methods, and sequential decision-making.

Predictive Analytics

Healthcare analytics, recommendation systems, and time-series forecasting.

Generative AI

Human motion synthesis, text generation, GANs, and diffusion models.

Computer Vision

Object detection, medical imaging, and visual understanding systems.


Technical Skills

Languages

Python R SQL C JavaScript HTML/CSS

ML & AI

PyTorch TensorFlow Hugging Face LangChain vLLM Unsloth TRL PySpark SKLearn

Agentic AI & LLM Tools

LangGraph LlamaIndex CrewAI AutoGen MCP OpenAI API Groq Streamlit

Infrastructure & Tools

AWS GCP Docker Kubernetes Kafka FastAPI Flask CI/CD

Databases & Vector Stores

MongoDB Redis Pinecone FAISS ChromaDB Elasticsearch

Let's Connect

I'm always excited to discuss new ideas, collaborate on research, or explore opportunities where I can make an impact. Feel free to reach out!