AI/GenAI Engineer (LLM Integration Specialist) _freelancer

SMARTnCODE Technology Private Limited; Hyderabad; On-site; INR 1,200,000 - 2,400,000; Full time

Job summary

SMARTnCODE Technology Private Limited is seeking an AI/GenAI Engineer to lead LLM integration, optimization, and domain customization for a high-performance chat application. You will connect multiple LLM APIs, implement robust wrappers, and design streaming real-time responses to support thousands of concurrent users.

You will collaborate with React and Python developers to create an intelligent chat experience, implementing prompt engineering, RAG pipelines, and cost-aware token usage while

Qualifications

• Experience working with LLMs and GenAI technologies.
• Proficiency with Python (FastAPI, LangChain, LlamaIndex preferred).
• Strong prompt engineering techniques and best practices.
• Experience with vector databases (Pinecone, Weaviate, Qdrant, ChromaDB).
• Knowledge of RAG implementation.
• Understanding of transformer architecture and attention mechanisms.
• Experience with API integration and streaming responses.
• Experience with LangChain, LlamaIndex, or similar LLM frameworks.
• Knowledge of fine-tuning techniques (LoRA, QLoRA, PEFT).
• Experience with embedding models and semantic search.

Responsibilities

• LLM Integration & Architecture: connect APIs, robust wrappers, streaming responses, rate limiting, token/context management.
• Domain Customization & Fine-tuning: domain prompts, RAG pipelines, LoRA/Prompt tuning, knowledge bases, few-shot learning.
• Performance & Optimization: reduce latency, caching, token-cost optimization, A/B testing, model routing pipelines.
• Safety & Quality: content moderation, guardrails, evaluation frameworks, hallucination handling, user feedback loops.
• Infrastructure: scalable LLM architecture, queuing, monitoring/logging, DevOps deployment for self-hosted models.

Skills

LLMs & GenAI

OpenAI API

Python

Prompt engineering

Vector databases

RAG implementation

Transformer architecture

API integration

LangChain / LlamaIndex

Fine-tuning (LoRA/QLoRA)

Embedding models

Multi-modal models

Tools

LangChain

LlamaIndex

FastAPI

vLLM

TGI (Text Generation Inference)

Job description

Role & responsibilities

About the Role

We're building a high-performance chat application an AI/GenAI Engineer to lead the integration and optimization of Large Language Models (LLMs). You'll be responsible for connecting LLM APIs, implementing domain-specific fine-tuning strategies, prompt engineering, and ensuring optimal performance for production use.

You'll work closely with our React and Python developers to create a seamless, intelligent chat experience that serves thousands of concurrent users.

Key Responsibilities

LLM Integration & Architecture

• Integrate multiple LLM APIs (OpenAI, Anthropic Claude, Google Gemini, or open-source models)
• Design and implement robust API wrapper services with retry logic, fallback mechanisms, and error handling
• Implement streaming responses for real-time chat experience
• Build rate limiting and quota management systems
• Handle token counting, context window management, and cost optimization

Domain Customization & Fine-tuning

• Develop domain-specific prompt engineering strategies
• Implement RAG (Retrieval Augmented Generation) pipelines using vector databases
• Fine-tune or adapt models for specific use cases using techniques like LoRA, prompt tuning
• Create and maintain knowledge bases for domain-specific responses
• Implement few-shot learning and in-context learning strategies

Performance & Optimization

• Optimize API response times and reduce latency
• Implement caching strategies for common queries
• Monitor and optimize token usage to control costs
• A/B test different models and prompts for quality improvements
• Build fallback chains (primary/secondary model routing)

Safety & Quality

• Implement content moderation and safety filters
• Build guardrails to prevent prompt injection and jailbreaking
• Develop evaluation frameworks to measure response quality
• Monitor and handle hallucinations and inaccuracies
• Implement user feedback loops for continuous improvement

Infrastructure

• Design scalable architecture for handling concurrent LLM requests
• Implement queue systems for managing high-volume API calls
• Set up monitoring and logging for LLM interactions
• Work with DevOps to deploy models (if self-hosted)

Preferred candidate profile

Must Have:

• Experience working with LLMs and GenAI technologies
• Strong experience with OpenAI API, Anthropic Claude, or similar LLM APIs
• Proficiency in Python (FastAPI, LangChain, LlamaIndex preferred)
• Strong understanding of prompt engineering techniques and best practices
• Experience with vector databases (Pinecone, Weaviate, Qdrant, ChromaDB)
• Knowledge of RAG (Retrieval Augmented Generation) implementation
• Understanding of transformer architecture and attention mechanisms
• Experience with API integration, webhooks, and streaming responses
• Strong problem-solving skills and ability to debug complex AI systems
• Experience with LangChain, LlamaIndex, or similar LLM frameworks
• Knowledge of fine-tuning techniques (LoRA, QLoRA, PEFT)
• Experience with embedding models and semantic search
• Familiarity with x library
• Experience deploying models using vLLM, TGI (Text Generation Inference)
• Knowledge of function calling/tool use with LLMs
• Experience with model evaluation metrics (BLEU, ROUGE, BERTScore)
• Understanding of token economics and cost optimisation
• Experience with open-source models (Llama, Mistral, Falcon)
• Knowledge of model quantization and optimization techniques
• Experience with multi-modal models (vision, audio)
• Familiarity with MLOps practices and experiment tracking (Weights & Biases, MLflow)
• Experience with AWS SageMaker, Google Vertex AI, or Azure ML
• Understanding of chain-of-thought prompting, ReAct, agents
• Experience building chatbots or conversational AI systems
• Publications or contributions to AI/ML community

Technical Stack You'll Work With

• Languages: Python (primary), JavaScript/TypeScript (basic understanding)
• LLM APIs: OpenAI, Anthropic, Google Gemini, Cohere
• Frameworks: LangChain, LlamaIndex, FastAPI
• Vector DBs: Pinecone, Weaviate, Qdrant, or ChromaDB
• Infrastructure: Docker, Redis, PostgreSQL, Message Queues
• Cloud: AWS/GCP/Azure (whatever your team uses)
• Monitoring: Prometheus, Grafana, custom LLM analytics