Razin (Saggad) M.
ML/AI Full Stack Engineer
Experience
Apr 2024 - Oct 2025
1 year 7 monthsML/AI Full Stack Engineer
AssistWell
- Deployed production CatBoost gradient boosting model for healthcare claim scoring, processing 100K+ claims/day with sub-second latency using singleton caching, optimized feature engineering and Pandas/NumPy pipelines.
- Developed PyTorch NLP pipeline with BERT transformers for automated ICD-10 medical code classification, fine-tuned on domain text achieving 90%+ accuracy and CPU-optimized inference.
- Deployed YOLOv8 models computer vision using PyTorch and MLflow for table structure detection in reimbursement forms, enabling real-time document inference and bounding box extraction.
- Built semantic search system with sentence transformers and ChromaDB, supporting vector-based similarity search and persistent/HTTP-managed collections for medical policy documents.
- Architected hybrid fraud detection system combining ML, rule-based and statistical algorithms with Django + Celery async processing, achieving 95%+ precision and PostgreSQL audit trails.
- Implemented explainable AI layer using case-based reasoning, retrieving similar cases from Parquet datasets and generating compliance-ready JSON outputs via Django REST Framework.
- Engineered large-scale ETL pipeline with Pandas multi-table joins, Pydantic data validation and PostgreSQL transformations handling complex healthcare claim hierarchies.
- Implemented distributed Celery/Redis async processing system with ACID transactions, error recovery mechanisms and Flower monitoring dashboard supporting 10K+ concurrent batch workflows.
- Established MLOps infrastructure with versioned CatBoost artifacts, multi-tenant deployments, PyTest evaluation frameworks and Azure Blob Storage metrics tracking for continuous model monitoring.
- Built Django REST Framework APIs with OpenAPI documentation, JWT/API key auth and Gunicorn WSGI server and CORS configuration serving 1M+ monthly requests at 99.9% uptime.
- Architected normalized PostgreSQL schema with 15+ Django ORM models, polymorphic storage, django-simple-history audit logs and 90% query optimization through ORM techniques.
- Built enterprise React/TypeScript application using Ant Design, TanStack React Query, styled-components, framer-motion and Vite build system with modular architecture.
- Implemented comprehensive testing with Vitest, Cypress E2E, Storybook, MSW mocking, ESLint/Prettier tooling and Sentry monitoring achieving 80%+ coverage.
May 2021 - Mar 2024
2 years 11 monthsGenerative AI & Backend Engineer
ColomboAI
- Engineered and deployed a generative AI API leveraging LLMs and RAG to automate marketing content creation, integrating LangChain, LangGraph and OpenAI models with AWS-based production infrastructure.
- Developed and integrated AI pipelines using LangChain and FastAPI to automate audio enhancement workflows, implementing asynchronous background task execution using Celery and AWS Lambda for scalable deployment.
- Built and deployed an AI-powered conversational chatbot trained on behavioral recovery literature, containerized with Docker and Kubernetes on GCP, leveraging LlamaIndex-based embeddings and reinforcement learning models to improve personalization and recommendation quality.
- Trained and optimized an AI assistant using RLHF and LoRA fine-tuning, improving NLP model accuracy and contextual response quality.
- Orchestrated large-scale AI automation workflows on AWS, using Step Functions, Bedrock, DynamoDB, Amplify, and Lambda to support real-time newsletter and media generation powered by multimodal LLMs including OpenAI and Claude.
- Built an end-to-end real estate valuation system with Airflow MLFlow and automating retraining and deployment for continuous predictions.
- Developed a natural language data-querying app with FastAPI and React, enabling Text-to-SQL generation and analytics through fine-tuned LLM prompts.
Feb 2020 - Apr 2021
1 year 3 monthsSoftware Engineer
Kustomer
- Built BERT and Bi-LSTM models for text classification, entity recognition, and chatbots, integrated with React/Next.js dashboards.
- Designed and deployed production AI models via Flask backend for content generation, ensuring robust API services, real-time inference, and fully tested integration with frontend applications.
- Developed interactive React/Next.js dashboards and web applications to visualize AI predictions and analytics, integrating seamlessly with Flask backend services and content generation pipelines.
- Engineered AWS-based ML pipelines with Docker, Lambda, and SageMaker, enabling automated inference and retraining workflows.
Jul 2016 - Jan 2020
3 years 7 monthsJunior Software Developer
Poweron Technology Services
- Built a real-time recommendation engine using collaborative filtering and scikit-learn, driving personalized user experiences.
- Developed embedding model for sentences splitting instead of word tokenizer using PyTorch, spaCy and Apache Kafka, improving operational efficiency.
- Implemented a computer vision application with OpenCV and TensorFlow, automating quality control processes in manufacturing.
- Created an automated machine learning pipeline with MLflow and scikit-learn, accelerating model development and deployment.
- Engineered a predictive maintenance system using time-series forecasting and Prophet, reducing downtime and maintenance costs.
- Conducted data visualization and exploratory data analysis (EDA) with Matplotlib and Seaborn, uncovering critical business insights.
- Collaborated with cross-functional teams to integrate ML models into existing systems, ensuring seamless operation and value delivery.
Summary
Senior ML/AI Full-Stack Engineer with 9+ years designing and deploying production-grade machine learning systems and enterprise applications. Expert in deep learning (PyTorch, TensorFlow), predictive analytics (CatBoost, XGBoost), and full-stack development (React/TypeScript, Python/Django, Node.js). Specialized in NLP, computer vision, and classical ML—building scalable AI solutions with Kafka, Celery, ChromaDB/Pinecone, AWS, Azure, MLflow, Docker, and Kubernetes.
Proven leader driving ML best practices, mentoring teams, and delivering measurable impact: $5M+ annual savings, 75% process automation, and 90%+ model accuracy across healthcare, fintech, and e-commerce domains.
Languages
English
NativeEducation
Oct 2012 - Jun 2016
Montana State University
Bachelor's Degree · Computer Science · United States
Need a freelancer? Find your match in seconds.
Try FRATCH GPT More actions
Similar Freelancers
Discover other experts with similar qualifications and experience
- English
- Deutsch
2025 © FRATCH.IO GmbH. All rights reserved.