Senior AI/ML Engineer — Applied AI

ACI Infotech; Hyderabad; On-site; INR 1,500,000 - 2,100,000; Full time

Job summary

ACI Infotech's Applied AI practice (ArqAI) helps enterprises move from AI strategy to production systems, delivering end-to-end AI solutions with a lead architect and delivery team across multiple industries.

Responsibilities include translating customer problems into applied-AI solutions, designing RAG pipelines and LLM features, owning model selection and production runbooks, building robust guardrails, and mentoring mid-level engineers on engagements.

Qualifications

• 5+ years in software or ML engineering, with 2+ years building and operating GenAI / LLM systems in production.
• Strong Python. Hands-on with LangChain, LlamaIndex, or LangGraph, and a vector database (Pinecone, Weaviate, pgvector, FAISS, or Qdrant).
• Production experience on cloud AI platforms (Azure OpenAI, AWS Bedrock, or Vertex AI) including deployment, cost, and security trade-offs.
• Working understanding of MLOps: experiment tracking, CI/CD for models, containerization, and observability for LLM apps.
• Comfort working with customer engineers and architects; able to run design reviews and document decisions.

Responsibilities

• Translate a customer problem statement into an applied-AI solution: what to retrieve, what to generate, what to evaluate, and what to leave to a deterministic system.
• Design and ship RAG pipelines, agentic workflows, and LLM-powered features deployed inside customer environments and integrated with their data, identity, and observability stack.
• Own model selection, prompt strategy, fine-tuning where it earns its keep, evaluation, and the production runbook.
• Build the guardrails — input filtering, output validation, evals, and observability — so the system is safe to operate after we leave.
• Lead technical reviews and mentor mid-level engineers on the engagement.

Skills

Python

LangChain

LlamaIndex

VectorDB

Tools

Azure OpenAI

AWS Bedrock

Vertex AI

Job description

ACI Infotech's Applied AI practice (ArqAI) takes enterprise customers from AI strategy to working production systems. Engagements run with a lead architect, a small delivery team, and a customer counterpart — across financial services, insurance, healthcare, retail, and manufacturing. Industries where uptime, auditability, and cost are non-negotiable.

Responsibilities

• Translate a customer problem statement into an applied-AI solution: what to retrieve, what to generate, what to evaluate, and what to leave to a deterministic system.
• Design and ship RAG pipelines, agentic workflows, and LLM-powered features deployed inside customer environments and integrated with their data, identity, and observability stack.
• Own model selection, prompt strategy, fine-tuning where it earns its keep, evaluation, and the production runbook.
• Build the guardrails — input filtering, output validation, evals, and observability — so the system is safe to operate after we leave.
• Lead technical reviews and mentor mid-level engineers on the engagement.

Requirements

• 5+ years in software or ML engineering, with 2+ years building and operating GenAI / LLM systems in production.
• Strong Python. Hands-on with LangChain, LlamaIndex, or LangGraph, and a vector database (Pinecone, Weaviate, pgvector, FAISS, or Qdrant).
• Production experience on Azure OpenAI, AWS Bedrock, or Vertex AI — including the deployment, cost, and security trade-offs.
• Working understanding of MLOps: experiment tracking, CI/CD for models, containerization, and observability for LLM applications.
• Comfort working with customer engineers and architects. You can run a design review, defend a technical decision, and write things down.

Nice to Have

• + Experience in a regulated industry (financial services, healthcare, insurance) and the compliance posture that comes with it.
• + Familiarity with MLflow, Weights & Biases, or similar tooling.
• + Background in fine-tuning, distillation, or running smaller open-weight models where the economics make sense.

Required Skills

LangChain LlamaIndex LangGraph vector database (Pinecone, Weaviate, pgvector, FAISS, or Qdrant)