Open to opportunities

Pranav Nijampurkar

@pranavnijampurkar

Message

ML Engineer specializing in LLMs, RAG, OCR, NLP and production-ready MLOps.

Japan

Message

What I'm looking for

I seek roles building production LLM/RAG systems with strong MLOps, collaborative teams, cloud-first infrastructure, mentorship opportunities and measurable business impact.

I am an ML Engineer with 7+ years of IT experience focused on machine learning, AI, and production MLOps. I specialize in LLMs, Retrieval-Augmented Generation, OCR, NLP, and deploying scalable AI systems on cloud platforms.

In my current role I built RAG+LLM chatbots and agentic AI solutions, cutting costs by 45% for a credit-card BPO and reducing invoice/receipt processing costs by 90% by moving from OCR to LLM-based approaches.

I have hands-on experience with model evaluation and orchestration across GPT-4, Claude, GeminiAI and LLaMA, and I employ tools like LangChain, LangFlow, MLflow, Weights & Biases, and vector databases (Milvus, Pinecone, Weaviate) for robust pipelines and A/B testing.

I bring full-stack experience in Python and Flask, cloud deployment expertise on GCP and Azure, plus bilingual Japanese capability (JLPT N2 / BJT-J3), enabling cross-functional collaboration and mentorship in AI/ML projects.

Experience

Work history, roles, and key accomplishments

Current

Senior ML Engineer

Current

UPSIDER Inc

Mar 2024 - Present (1 year 7 months)

Led development of an AI-powered RAG chatbot and agentic AI solutions, reducing credit-card BPO operational costs by 45% and cutting receipt/invoice processing costs by 90% while improving accuracy.

Python Flask LLMs Milvus MLOps GCP LLM

Data Scientist / MLOps Engineer

KPMG Ignition Tokyo

Dec 2021 - Feb 2024 (2 years 2 months)

Designed document parsing and RAG-based pipelines using OCR, Azure Form Recognizer and OpenAI APIs, built CI/CD for ML deployments and led client onboarding of Azure resources for scalable document-processing solutions.

Python Django NLP RAG MLFlow CI CD Docker

Data Scientist (Computer Vision)

Persol Technology

Nov 2019 - Nov 2021 (2 years)

Developed and optimized behavioral detection systems using CNNs, YOLO/SSD and OpenPose; researched template matching and monodepth estimation for fisheye cameras to improve detection accuracy in real-world deployments.

Python TensorFlow Keras OpenCV YOLO SSD Transfer Learning Faiss AWS

Full Stack Developer / Data Analyst

Harbinger Systems

Jul 2017 - Sep 2019 (2 years 2 months)

Built e-learning and job-posting systems, implemented web APIs and automation, and prototyped chatbots integrated with OCR and LSTM models to streamline content and job-post workflows.