Hamza Khan

Lead Machine Learning Engineer/Generative AI Lead

Horizon Tech

May 2021 - Current (5 years 3 months)

Led the design and deployment of LLM-powered agents using a LangChain-based orchestration, integrated into React-based enterprise workflows. Built a retrieval-augmented generation (RAG) system on Contextual AI's proprietary vector search service using OpenSearch and FAISS for improved context relevance. Established evaluation protocols (BLEU, ROUGE, hallucination detection) to monitor and validate generative output quality. Conducted fine-tuning of Llama-style and GPT-J models on domain-specific datasets to tailor them for business use cases. Implemented CI/CD pipelines for model serving via AWS Lambda, ECS, and managed SageMaker, enabling seamless deployment and version control.

R

Senior Machine Learning Engineer

ResolveAI Systems

Feb 2018 - Apr 2021 (3 years 3 months)

Customized and trained transformer models (BERT, ROBERTa) using Hugging Face for applications such as summarization, semantic search, and legal document processing. Built end-to-end ML pipelines: data preprocessing, training, serving, and monitoring, using MLOps tooling built around Seldon Core. Created bias, toxicity, and factuality monitoring dashboards for generative AI outputs to help ensure ethical use in production. Automated retraining workflows, enabling continuous model improvement and versioning through MLflow. Mentored engineers in best practices for production ML, from modular architecture to code reviews and reproducible training.

A

Machine Learning/NLP Engineer

AI21 Labs

Aug 2016 - Jan 2018 (1 year 6 months)

Designed NLP pipelines for named entity recognition, topic modeling, and sentiment analysis using spaCy, NLTK, and Gensim. Developed RESTful APIs (FastAPI, Flask) to serve trained models with low latency, integrated with client applications. Collaborated closely with domain experts to validate model outputs, ensuring business relevance and accuracy. Built monitoring systems to track drift and performance degradation in production, alerting when model metrics changed.

Data Scientist

Primer

Feb 2014 - Dec 2016 (2 years 11 months)

Built predictive models (classification, regression) and NLP solutions for clients across industries, helping automate decision processes. Developed complete ETL pipelines: data extraction, validation, feature engineering, and scalable training infrastructure. Optimized model performance through hyperparameter tuning and cross-validation, increasing accuracy and reducing overfitting. Partnered with business stakeholders to translate insights into dashboards and visual reports for decision-making.

About

Detail

Timeline

Résumé