ML Engineer
2+ years across the ML lifecycle — document AI, predictive work, and production inference.
Hi, I'm Zaheer K Z
About
Applied AI with product-level polish.
The story matches the interface: strong engineering underneath, calm presentation on top — ML delivery, disciplined execution, and UI quality in one place.
Experience
Neutrinos · ML Engineer II · Feb 2024 – Present · Bengaluru
Education
BCA · University of Mysore · ML at Brototype (Calicut).
Craft
Building scalable AI pipelines, agent-style workflows, and microservices that stay maintainable after launch.
I care how cutting-edge ML shows up in the real world — from RAG systems to dependable deployments on Kubernetes.
PythonFastAPIMongoDBDockerPyTorchTensorFlowKubernetesRabbitMQRedisCeleryAWSGitMLflowTransformersLangChainOpenCVRAGNLPPythonFastAPIMongoDBDockerPyTorchTensorFlowKubernetesRabbitMQRedisCeleryAWSGitMLflowTransformersLangChainOpenCVRAGNLP
Open to new opportunities & collaboration
Bengaluru, India
Map: Wikimedia Commons
Mindset
Building more than models. Side projects and football give me the discipline and focus to keep shipping.
Mastering the craft — from training models to autonomous agent systems — is how I pursue excellence.
Football
Weekend five-a-side leagues
Projects
Applied ML, document intelligence, and research tools—built end to end.
01
ML Platform
AI Orchestrator
Orchestration for document AutoML pipelines, queue-backed FastAPI services, MongoDB, and GenAI features—PII masking, OCR, RAG, MCP—containerized on Docker and Kubernetes.
02
Document AI
Document Key-value Pair Extraction
Key–value field detection and extraction with YOLOv8, OCR, and inference served through a dedicated API—field-level accuracy up to roughly 99% on structured document workloads.
03
NLP Platform
PaperMate
Personalized arXiv recommendations in AI, ML, DL, and CV; summarization and paper Q&A—with reported user satisfaction from about 65% to about 85%.
Tech Stack
My Skills
Loading 3D visualization...
Testimonials
What they say
1,200+
Reviews
10,000+
Happy clients
The GenAI guardrail work was a game-changer for our team—PII masking in production and a RAG pipeline that actually cut response times.
Rare mix of LangChain depth and FastAPI discipline. The platform scaled from dozens to thousands of daily requests without drama.
Document extraction jumped in accuracy after the YOLOv8 + OCR pass. Clean code we could own and extend.
Mentoring was practical and clear—hard ML topics landed fast for the whole cohort.
Shipped MLflow + K8s patterns we had been circling for weeks. Four days from brief to working pipeline.
Chroma + Redis caching cut semantic-search latency sharply—exactly the performance lens we needed.
The GenAI guardrail work was a game-changer for our team—PII masking in production and a RAG pipeline that actually cut response times.
Rare mix of LangChain depth and FastAPI discipline. The platform scaled from dozens to thousands of daily requests without drama.
Document extraction jumped in accuracy after the YOLOv8 + OCR pass. Clean code we could own and extend.
Mentoring was practical and clear—hard ML topics landed fast for the whole cohort.
Shipped MLflow + K8s patterns we had been circling for weeks. Four days from brief to working pipeline.
Chroma + Redis caching cut semantic-search latency sharply—exactly the performance lens we needed.
Other
More ways to explore the profile.
Resume Snapshot
A local copy of the resume is included so the portfolio stays self-contained.
Open ResumeAchievements
98% precision on the classifier project, production AI delivery, and ML-focused academic coursework.
See WorkAvailability
Open to AI product work, applied ML delivery, and high-impact collaboration opportunities.
Contact