Predictive Modeling of Skin Permeability
Co-author of research paper on Predictive Modeling of Skin Permeability for Molecules. Published in PLOS Digital Health journal with 30 citations.
Read PaperI build AI systems that cut costs, automate decisions, and ship to production fast.
Senior Data Scientist & ML Engineer · 4× AWS Certified
Replace manual, repetitive workflows with AI that works 24/7 — cutting labor costs and eliminating human error at scale.
Demand forecasting, anomaly detection, churn prediction, recommendation systems — turn your historical data into a system that keeps you one step ahead.
From AI agents and multi-agent systems to full production deployments — monitored, CI/CD-backed, and ready in weeks. No black-box handoffs.
Python, SQL, TypeScript, Node.js, R, Bash (Linux), FastAPI, PyDantic, Streamlit, REST APIs, GraphQL APIs, OOP, Git
Scikit-learn, XGBoost, LightGBM, PyTorch, TensorFlow, Computer Vision, HuggingFace, Transformers, RAG, LLMs, LangChain, LangGraph, LangSmith, LangFuse, Opik, CrewAI, n8n, vLLM, LoRA, QLoRA, Quantization, MCP, Multi-Agent Systems, LLM-as-a-Judge, Hybrid Search, Reranking, Embeddings Pipelines
MLFlow, AWS SageMaker, Docker, GitHub Actions, CI/CD Pipelines, Observability & Tracing, LLMOps, LangFuse, Opik, DVC, Container Orchestration
PostgreSQL, MongoDB, Redis, Pinecone, Qdrant, FAISS, Apache Airflow, Apache Spark, PySpark, Dagster, Inngest, Data Modeling, ETL/ELT Pipelines
AWS (S3, Glue, Athena, Lambda, SQS, SNS, ECR, ECS, ELB, RDS, DynamoDB, ElastiCache, Bedrock, SageMaker) · GCP (Cloud Run, Vertex AI, Gemini)
NumPy, Pandas, Polars, Matplotlib, Seaborn, Power BI, Hypothesis Testing, A/B Testing, Data Mining, Statistical Analysis
End-to-end ML pipeline predicting NYC taxi trip durations with automated CI/CD workflows, deployed as serverless REST API using AWS Lambda and API Gateway.
View ProjectFine-tuned Qwen2.5-1.5B-Instruct and LLaMA 3.1-8B using Unsloth and LoRA on custom instruction-style dataset for efficient, low-resource training.
View ProjectProduction-ready RAG-based chatbot for intelligent document Q&A over PDFs, powered by gpt-4o-mini and LangChain with Arabic and English support.
View ProjectUtilizes Gemini API with gemini-1.5-flash model to enhance resumes, generating insights like strengths, weaknesses, job fit assessment, and skills gap analysis.
View ProjectJune 2025
July 2024
Co-author of research paper on Predictive Modeling of Skin Permeability for Molecules. Published in PLOS Digital Health journal with 30 citations.
Read PaperA cautionary tale about an SQS-Lambda integration gone wrong. Learn how a missing Dead Letter Queue and a 200-page PDF triggered an infinite retry loop that ran for 2 days, costing $1,500. The first step toward building a robust, production-grade document processing system.
Read MoreLet's connect! Schedule a meeting to discuss collaboration opportunities, projects, or just to chat about data science and machine learning.