Looking for a job

Latasha Nayak

@latashanayak

Machine Learning Intern at Hippo Video AI who built RAG systems that cut hotel RFP response time from 3 hours to 30 minutes.

India

Message

At Hippo Video AI, I engineered two LangChain RAG pipelines for hotel RFP responses, reducing manual effort by 80% and cutting response time from three hours to 30 minutes. I also developed an FAQ system with 1,800+ question-and-answer pairs that achieved up to 95% retrieval precision.

As an Undergraduate Research Assistant at NIT Trichy, I fine-tuned a Llama 3.1 8B instruct model on MITRE ATT&CK ICS data and built a RAG pipeline for secure local inference. The work achieved 92% accuracy over the baseline model.

My projects include TrendSense, which clusters news articles to identify emerging trends, and a machine unlearning system that reduced model retraining time by 50%. I’m studying Instrumentation and Control Engineering at National Institute of Technology Tiruchirappalli.

Experience

Work history, roles, and key accomplishments

Undergraduate Research Assistant

NIT Trichy

May 2026 - Jul 2026 (2 months)

Fine-tuned a Llama 3.1 8B instruct model using LoRA/PEFT on 4,900+ MITRE ATT&CK ICS STIX-derived samples for automated ICS threat modelling. Built an end-to-end RAG pipeline with LangChain, ChromaDB/FAISS, and Ollama for secure local inference on ICS cybersecurity knowledge.

Python PyTorch PEFT RAG Lang Chain ChromaDB Faiss Ollama MITRE ATT&CK STIX NLP

Machine Learning Intern

Hippo Video AI

May 2025 - Jul 2025 (2 months)

Engineered 2 RAG pipelines using LangChain to automate hotel RFP responses, reducing manual effort by 80% and processing RFPs 6x faster. Designed Regex-based parsers handling 300+ RFP documents/month and developed an FAQ RAG system with 1800+ Q&A pairs achieving sub-3s latency and up to 95% retrieval precision.

Python Lang Chain RAG Regex JSON FAQ Systems Retrieval NLP