Henry Abdullah

AI/ML Engineer with 6+ years of experience building and deploying Machine Learning, NLP, and Generative AI solutions across cloud and enterprise environments. Skilled in developing end-to-end AI systems, including predictive models, intelligent search platforms, Retrieval-Augmented Generation (RAG) applications, and LLM-powered assistants using Python, PyTorch, FastAPI, LangChain, vector databases, and AWS. Proven track record of delivering scalable, production-grade AI solutions that improve operational efficiency, decision-making, and user experience.

AI/ML Engineer at Istream Solutions (2023-01 – Present)

• Built enterprise AI-powered SaaS knowledge systems using LLMs, OpenAI APIs, LangChain, and RAG to enable intelligent document search and automation.
• Developed enterprise semantic search platforms using LangChain, LlamaIndex, Pinecone, FAISS, and Elasticsearch for large-scale knowledge retrieval.
• Fine-tuned transformer models using Hugging Face and PyTorch for improving enterprise NLP and GenAI application performance.
• Built vector-based search systems using embeddings, FAISS, and Pinecone to support high-scale similarity search in SaaS platforms.
• Designed AI assistants and workflow automation tools using LLMs and agent-based systems to improve business process efficiency in enterprise environments.
• Developed production-grade ML APIs using FastAPI, Docker, AWS, and cloud infrastructure for scalable SaaS deployment.
• Integrated ML models into SaaS platforms for intelligent search, document understanding, and workflow automation features.
• Implemented model evaluation and monitoring pipelines using MLflow and metrics tracking to ensure production reliability of AI systems

Machine Learning Engineer at Kintsugi (2020-05 – 2022-12)

• Worked on healthcare and clinical data analytics systems using Python, SQL, Pandas, NumPy, and PostgreSQL for preprocessing and feature engineering.
• Built NLP pipelines for clinical text analysis and EHR-style data processing using spaCy, NLTK, and transformer-based models.
• Applied Named Entity Recognition (NER) and semantic NLP techniques to extract medical entities from unstructured healthcare records.
• Developed predictive models for mental health risk assessment and behavioral pattern analysis using machine learning techniques.
• Evaluated ML models using cross-validation and performance metrics to improve clinical NLP system reliability and accuracy.
• Deployed and maintained ML inference APIs using FastAPI, Docker, and AWS for secure healthcare data processing systems.

Bachelor of Science in Computer Science