LLM Engineer

LLM Engineer

Design, build, optimize, deploy, and operate Large Language Model (LLM) and Small Language Model (SLM) capabilities.

Build secure, reliable, reusable, and enterprise-ready AI capabilities.

Support:
Agentic AI workflows, AI for SDLC, Knowledge retrieval, Model evaluation, Private AI hosting, Agent Ops.

Work closely with:
Principal AI Architect, AI Engineering Lead, Platform Engineers, Security teams, Enterprise Architecture, Product Owners, Domain teams.

Must-Have Technical Skills

• Large Language Models (LLMs)
• Small Language Models (SLMs)
• Prompt Engineering
• Context Engineering
• Retrieval-Augmented Generation (RAG)
• Embeddings
• Semantic Search
• Agentic AI Patterns
• Multi-Agent Workflows
• Tool Calling
• Function Calling
• Model Evaluation
• LLM Observability

• Fine-Tuning
• Supervised Fine-Tuning
• LoRA
• QLoRA
• Quantization
• Distillation
• Model Compression
• Synthetic Data Generation
• Model Benchmarking
• Model Selection
• Model Routing

• Private LLM Hosting
• On-Prem Model Deployment
• GPU-Based Inference
• Model Serving APIs
• High-Availability Inference
• Autoscaling
• Load Balancing
• Caching
• Batch and Real-Time Inference

• Kubernetes
• Docker
• Kubeflow
• KServe
• Ray Serve
• MLflow
• Hugging Face
• Transformers
• Py Torch
• PEFT
• Deep Speed
• NVIDIA NIM
• Triton Inference Server
• TensorRT-LLM
• vLLM
• TGI
• SGLang

• Python
• Type Script or Java Script
• REST APIs
• Microservices
• CI/CD
• Git Hub or Azure Dev Ops
• API Design
• Distributed Systems
• Cloud-Native Engineering
• Test Automation

• Vector Databases
• Knowledge Graphs
• Document Processing
• Metadata Management
• Data Pipelines
• Object Storage
• Enterprise Search
• Structured and Unstructured Data Integration

• Build enterprise-grade LLM-powered applications and intelligent agent capabilities.
• Design reusable LLM patterns, services, APIs, and accelerators.
• Develop model interaction patterns for:
  Reasoning, Summarization, Classification, Extraction, Planning, Decision support.
• Build reusable prompt, context, retrieval, memory, and evaluation components.
• Support AI-for-SDLC agents across:
  Requirements, Design, Coding, Testing, Security Review, Deployment, Operations.
• Convert AI use cases into scalable production solutions.

• Build core intelligence services for enterprise agents.
• Develop reusable capabilities for:
  Planning, Task decomposition, Reasoning, Tool usage, Agent collaboration.
• Enable agent-to-agent interaction and multi-agent orchestration.
• Integrate LLMs with:
  Agent runtimes, Tool registries, Workflow engines, MCP-based gateways.
• Support human-in-the-loop, approval, escalation, and feedback workflows.
• Improve agent quality, accuracy, safety, and task completion.

• Design reusable prompt engineering standards, templates, and libraries.
• Create:
  System prompts, Task prompts, Role prompts, Guardrail prompts, Evaluation prompts.
• Develop context engineering strategies for better grounding, relevance, and personalization.
• Optimize:
  Token usage, Context windows, Memory injection, Retrieval inputs.
• Establish prompt versioning, testing, and governance practices.

• Design and implement enterprise RAG architectures.
• Build retrieval pipelines using:
  Enterprise documents, Knowledge repositories, Structured data, Metadata.
• Optimize:
  Chunking, Embeddings, Indexing, Ranking, Reranking, Retrieval strategies.
• Improve grounding, citation quality, precision, recall, and factual accuracy.
• Build reusable retrieval services for agents and business domains.
• Partner with data and knowledge management teams to onboard trusted data sources.

• Evaluate, build, fine-tune, deploy, and optimize LLMs and SLMs.
• Support domain-specific model development using approved datasets.
• Build supervised fine-tuning and model adaptation pipelines.
• Apply:
  LoRA, QLoRA, Distillation, Quantization, Model compression.
• Evaluate commercial, open-source, and internally hosted models.
• Select models based on:
  Accuracy, Latency, Cost, Data residency, Security, Operational requirements.

• Build and support private AI capabilities for LLM/SLM hosting.
• Deploy models across:
  On-premises, Hybrid, Private cloud environments.
• Support GPU-enabled model hosting.
• Optimize latency, throughput, concurrency, resiliency, and GPU utilization.
• Build secure inference endpoints for internal applications and agents.
• Support air-gapped and restricted AI environments.
• Partner with infrastructure and platform teams on private AI hosting.

• Implement scalable model serving using modern inference frameworks.
• Build high-availability inference architectures.
• Optimize:
  Token throughput, Response latency, Cost efficiency, Inference performance.
• Implement:
  Model routing, Load balancing, Caching, Fallback strategies.
• Support batch and real-time inference.
• Develop reusable deployment templates for different model families.

• Build operational…