Looking for a job

Amay Dubey

@amaydubey

Senior Machine Learning Engineer at M-Files, where I built an enterprise retrieval and ranking system achieving 85%+ result relevance.

India

Message

At M-Files, I architected and productionized an enterprise semantic retrieval and ranking system that achieved 85%+ result relevance across thousands of users. I also designed a RAG platform and evaluation frameworks that improved query-resolution efficiency by 85%.

At M-Files, I developed document intelligence and GraphRAG workflows, including metadata extraction that reached 95% accuracy and reduced manual entry effort by 90%. Earlier, at Prescinto, I built an analytics assistant for solar-plant and BESS data, and at Paxcom I reduced CPU inference time by 85% using ONNX Runtime.

Experience

Work history, roles, and key accomplishments

Current

Senior Machine Learning Engineer

Current

M-Files

Feb 2025 - Present (1 year 9 months)

Architected and productionized an in-house semantic retrieval and ranking system, achieving 85%+ result relevance across thousands of enterprise users. Designed and deployed a scalable RAG platform and led document intelligence and agentic AI initiatives.

Python PyTorch RAG LangGraph VLLM ONNX Runtime Docker Kubernetes Azure Graph RAG LLM Evaluation PEFT LoRAs QLoRA MCP Copilot Extensibility

Senior Data Scientist

Stealth AI Startup

Jul 2024 - Nov 2024 (4 months)

Built an automated unstructured-to-structured data pipeline combining NLP, OpenAI models, and fuzzy matching, improving cross-dataset consistency by 90%. Engineered a multiclass Random Forest classifier and a transformer-based NER pipeline achieving 95% accuracy.

Python NLP OpenAI Random Forest SpaCy NER Fuzzy Matching

Data Scientist

Prescinto

Aug 2022 - Jun 2024 (1 year 10 months)

Fine-tuned open-source LLMs for text-to-SQL generation and built an LLM-powered analytics assistant for solar-plant and BESS data. Developed predictive models for battery health and warranty expiry.

LLM Fine tuning PEFT LoRAs QLoRA Text To SQL Prompt Engineering Agentic AI ML Pipelines predictive modeling

Data Scientist

Paxcom

Feb 2022 - Jul 2022 (5 months)

Developed aspect-based sentiment analysis models for e-commerce clients and reduced CPU inference time by 85% by converting models to ONNX Runtime. Built Flask REST APIs for model serving.

NLP Sentiment Analysis ONNX Runtime Flask REST APIs

Senior Software Engineer

Capgemini

Jun 2019 - Feb 2022 (2 years 8 months)

Developed an entity-resolution and data-deduplication solution for MetLife Insurance and converted Informatica mappings into a Neo4j knowledge graph. Led LLM proof-of-concepts for legal-text summarization.

Entity Resolution Data Quality Neo4j Knowledge Graphs LLM Transformers