AI Infra Engineer

About the Role

We are looking for a hands-on AI Infrastructure Engineer to set up and operate open-source AI models on local hardware and develop Python-based automation for document and video data. The consultant will build reliable offline processing workflows that can be run independently by a nontechnical team member, with an emphasis on confidentiality, accuracy, and ease of use.

Key Responsibilities

· Install, configure, and troubleshoot open-source AI models and inference tools on local computers, without relying on cloud-based AI services.

· Develop Python scripts to extract and process text from PDFs, scanned documents, and screen recordings; integrate local OCR or transcription tools where needed.

· Build automated workflows for file inventory, document organization, duplicate detection, and structured output generation.

· Generate traceable processing outputs, descriptive filenames, metadata, and clear status/error reports.

· Protect original files and confidential data by using offline processing, separate output folders, and secure handling practices.

· Implement reliable error handling and restartable processing so interrupted or repeated runs do not lose work or create duplicate outputs.

· Document setup and operation clearly, and hand over easy-to-run scripts to nontechnical users.

Required Qualifications

· Strong practical Python programming and automation skills.

· Demonstrated experience running open-source AI models locally using tools such as Ollama, vLLM, llama.cpp, or Hugging Face.

· Hands-on familiarity with models such as Llama, Qwen, Mistral, or DeepSeek.

· Experience extracting text from documents using PDF-processing and OCR tools such as PyMuPDF, Tesseract, or PaddleOCR.

· Understanding of offline data processing, file handling, logging, and confidentiality safeguards.

· Ability to work independently, troubleshoot technical issues, and explain the solution to nontechnical users.

· Based in Pune / Pimpri-Chinchwad, with availability for approximately 5–10 hours per week and occasional on-site work.

Nice to Have

· Experience with local audio/video transcription tools such as Whisper.

· Knowledge of Docker, NVIDIA GPUs, CUDA, and GPU-based model deployment.

· Exposure to RAG, LangChain, or LlamaIndex.

· Experience with large document archives, SHA-256 hashing, near-duplicate detection, JSON metadata, and restartable processing pipelines.

Join us at DIGIPATH AI and contribute your technical expertise to building reliable, practical AI solutions.

Pay: Up to ₹80,000.00 per year

Work Location: Hybrid remote in Pune, Maharashtra