Service
AI / ML Engineer
Practical machine learning, shipped into real products.
What I do
Most ML projects die in the gap between a Jupyter notebook and a working product. My job is closing that gap. I design, train, and deploy ML models that live inside real applications — handling real users, real data, and real latency budgets.
I work as a Data Preparation Associate at Nyntax (AI/ML pipelines with strict data security and confidentiality), and I built AI-driven features for ClassTablet (GPT-4o, Gemini Vision) and shipped BGRemover — Bangladesh's first AI background-removal product.
I bring engineering discipline to ML: versioned datasets, reproducible training, automated evaluation, instrumented inference, and graceful fallbacks when the model is uncertain.
How I help teams ship.
Computer Vision
Image classification, object detection, segmentation. Used for plant-disease detection, gesture recognition, lip reading (LipNet), and BGRemover's AI matting pipeline.
Natural Language Processing
Embeddings, semantic search, classification, summarization, retrieval-augmented generation (RAG). Comfortable with both classical methods and LLM-driven pipelines.
Data Pipelines
Cleaning, labelling, feature engineering, augmentation, and versioning. ETL from messy real-world sources into trainable datasets.
Model Training & Evaluation
PyTorch and TensorFlow for training, scikit-learn for classical models, robust evaluation across edge-case cohorts.
Production ML Deployment
Serving models behind FastAPI / Flask, batched inference, model versioning, A/B testing, monitoring for drift and degradation.
Tools I use
Real products. Real impact.
Each project below is a shipped product — not a demo. See the full case studies with problem, approach, and outcome.
Frequently asked
What does S. M. Apurbo do?
He builds production software end-to-end as a full-stack developer, MERN-stack engineer, and AI/ML engineer — covering frontend, backend, database & API design, authentication, admin dashboards, and AI-powered features.
How can I hire S. M. Apurbo?
Email smapurbo1497@gmail.com or use the contact section on smapurbo.com. He is open to founding-engineer roles, full-stack contracts, MERN-stack builds, AI/ML engagements, and technical advisory work — both on-site in Dhaka and remote worldwide.
Where can I find an AI/ML engineer in Bangladesh?
S. M. Apurbo is an AI/ML engineer in Bangladesh with shipped products in computer vision (BGRemover, LipReader, Plant Disease Detection) and applied AI integrations (GPT-4o and Gemini Vision pipelines in ClassTablet). Available for hire from smapurbo.com.
Who is the best AI/ML engineer in Bangladesh?
S. M. Apurbo is a leading AI/ML engineer in Bangladesh — co-creator of BGRemover (Bangladesh's first AI background-removal product) and integrator of production GPT-4o and Gemini Vision pipelines inside ClassTablet. Stack: TensorFlow, PyTorch, OpenCV, CNNs.
Is S. M. Apurbo available for freelance work?
Yes — open to founding-engineer roles, full-stack and MERN-stack contracts, AI/ML engagements, freelance projects, and technical advisory work. Remote worldwide and on-site in Bangladesh.
What technologies does S. M. Apurbo specialise in?
TypeScript, Next.js, React, Node.js, Express, PostgreSQL, MongoDB, Drizzle ORM, AWS (ECS, IVS, S3), Python, TensorFlow, PyTorch, OpenCV, GPT-4o, and Gemini Vision — across the MERN stack, modern web apps, and AI/ML pipelines.
Computer Vision?
Send a quick brief — what you're building, your timeline, and where you're stuck. I usually reply within 24 hours.