Looking for a job

Hosna Hatami

@hosnahatami

AI/ML Engineer focused on NLP, LLMs, RAG, and applied AI. I enjoy building practical, reliable AI systems and solving real-world problems with ML.

Iran

Message

What I'm looking for

I'm looking for remote or relocation opportunities where I can build production AI systems end to end, especially generative AI, NLP, computer vision, and machine-learning products.

I build applied AI products end to end, from data preparation and fine-tuning through deployment on AWS, GCP, and Firebase. At Faranab Energy, I develop plant-identification and disease-detection models, an in-app LLM assistant, and ML/NLP-powered recommendation systems.

At Creative Intell, I built a legal-domain LLM assistant for a New York team, including hybrid RAG across 45,000+ pages of legal documents. I reduced latency from roughly 17 seconds to under one second and lowered hallucinations from 12% to under 3% through routing, retrieval, verification, and human review.

At Kara Group, I worked across medical imaging, clinical NLP, real-estate analytics, and financial analytics. I also owned MLOps workflows for containerized training, deployment, CI/CD, and model monitoring.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer

Current

Faranab Energy

Apr 2026 - Present (5 months)

Building applied AI for consumer products, including computer vision models for plant identification and disease detection, an LLM-powered assistant with function calling, and recommendation systems. Owning the stack on Firebase, GCP, and AWS with model serving, data pipelines, CI/CD, and monitoring.

Computer Vision Disease Detection LLM Function Calling Recommendation Systems NLP Firebase GCP AWS CI CD Monitoring

AI Engineer

Creative Intell

May 2025 - Apr 2026 (11 months)

Designed and built a legal-domain LLM assistant end to end, including architecture, model selection, retrieval integration, and evaluation. Engineered a RAG pipeline with hybrid retrieval over 45,000+ pages of legal documents, reducing latency from 17 seconds to under one second and cutting hallucination rate from 12% to under 3%.

LLM RAG Faiss Weaviate ChromaDB Lang Chain LangGraph LoRAs QLoRA DeepSpeed Sagemaker Vertex AI Mixture Of Experts Caching Parallel Retrieval Semantic Chunking Citation Verification Confidence Scoring Human In The Loop

AI Engineer

Kara Group

Jun 2023 - May 2025 (1 year 11 months)

Developed medical imaging models for breast-cancer detection and clinical NLP tools, and led analytics projects in real estate and finance. Owned the MLOps stack including CI/CD and containerized deployment.