Forward Deploy AI Engineer
On-site · Bengaluru, Karnataka, India
Job Summary
Staff Data Scientist – Supply Chain AI (Small Language Models) in India. You will design, fine-tune, and deploy optimized Small Language Models that power production-grade AI features across logistics, planning, inventory, and demand forecasting. Build and fine-tune SLMs/LLMs for production use cases such as semantic search, forecasting, contract intelligence, and conversational insights; optimize models for performance, cost, and latency using PEFT, quantization, and efficient inference techniques. Embed supply chain knowledge into models by collaborating with domain experts and product teams. Enable agentic workflows where AI Assistants execute planning, optimization, and decision-support tasks. Integrate models into products using a Graph-based microservices architecture and custom Lambdas. Apply vector search and semantic reasoning to navigate complex supply chain relationships. Work with ERP, logistics, inventory, supplier, and contract datasets; deploy and operate models with strong MLOps practices on AWS, GCP, or Azure. Measure and improve quality to reduce hallucinations and improve reliability through continuous evaluation. Partner with engineering, product, and operations teams to ship impactful AI features end-to-end.
Required Qualifications
- Experience working with small language models (SLMs) and LLMs (e.g., Mistral, Phi, Nemotron, Llama variants) including parameter-efficient fine-tuning (LoRA/QLoRA) for production-ready QA or knowledge tasks
- Strong understanding of supply chain data and workflows (logistics, inventory, procurement, ERP systems)
- Familiarity with Knowledge Graphs and graph-based data modeling, including designing and querying Neo4j graphs
- Solid hands-on skills in Python, PyTorch/TensorFlow, GraphQL, and vector databases
- 8+ years of experience in forecasting, optimization, supply chain analytics, or AI-driven products
- Experience designing Retrieval-Augmented Generation pipelines with document chunking, metadata filtering, and hybrid (dense + sparse) retrieval for enterprise knowledge search
- Hands-on experience building and tuning FAISS-based vector indexes for large-scale semantic search with low-latency production deployment
- Experience taking models from prototype to production
- Clear communication skills and strong product ownership
Apply with one swipe on Sorce. We auto-fill applications and apply on your behalf — no cover letters, no 40-minute forms.
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.