Confirmed on the employer's own hiring board on Nov 20, 2026. First seen by Alion on Nov 20, 2026. Technource scores B on the Alion truth index.
Qualifications
Build, train and fine-tune custom ML/DL models for NLP, computer vision and prediction tasks (classification, regression, clustering, recommendation)
Handle the full model process: collect and clean data, train, test, deploy and monitor models
Fine-tune open-source LLMs (Llama, Mistral, etc.) using LoRA/QLoRA on client data, and build RAG systems with vector databases when needed
Strong Python and PyTorch: write clean code and build your own model architectures and training loops, not just use ready-made APIs
Clear basics of how ML works: neural networks (backpropagation, loss functions, optimisers), Transformers (attention, embeddings, tokenisation) and the maths behind them
Computer vision and deployment: hands-on with OpenCV, YOLO, face recognition and OCR; deploy models using FastAPI, Docker and ONNX. AWS and MLOps tools are a plus
Good English communication and presentation: build proofs of concept and clearly explain your models, results and timelines to clients and the team
Education: B.E./B.Tech (CS, IT, AI or related) or M.Tech/M.Sc in CS, AI, Data Science or Mathematics

