21 Nov
|
Abh Consultants
|
India
21 Nov
Abh Consultants
India
Senior Generative AI & LLM Engineer — Freelance / Part-Time Technical Support
Job Type: Freelance / Part-Time / Contract
Engagement: Approximately 10–15 hours per weekDuration: Ongoing, based on project requirementsWork Mode: Remote (Pan-India)
Experience Required: 6–9 years
Opportunity: Suitable for experienced Generative AI Engineers, LLM Engineers, AI Architects, and Machine Learning Engineers seeking additional freelance income or part-time consulting opportunities alongside their existing employment.
We are seeking experienced professionals with strong hands-on expertise in Python, PyTorch, LangChain, LlamaIndex, OpenAI/Anthropic APIs, Retrieval-Augmented Generation (RAG), Vector Databases, Agentic AI, LLM Fine-Tuning, and Production-Scale Generative AI Applications.
1. Core Generative AI & LLM Technologies
- Python – Advanced programming, asynchronous processing, reusable libraries, API development, and AI application engineering.
- PyTorch – Deep learning model development, neural networks, model training, and GPU-based computation.
- LangChain – LLM application development, chains, tools, memory management, and agent integrations.
- LlamaIndex – Data ingestion, indexing, retrieval, and knowledge-based LLM application development.
- OpenAI / Anthropic APIs – GPT models, Claude models, prompt engineering, function calling, and structured outputs.
- Hugging Face Transformers – Pre-trained models, tokenizers, model deployment, and fine-tuning workflows.
- FastAPI – Building scalable REST APIs, asynchronous AI services, and LLM-powered backend applications.
- RAG Pipelines – Document processing, chunking, embeddings, retrieval, context management, and response generation.
2. LLM Architecture, Agentic AI & Fine-Tuning
- Agentic AI Workflows – Designing autonomous AI agents capable of reasoning, planning, tool usage, and task execution.
- Multi-Agent Orchestration – Coordinating multiple specialized agents using LangGraph, CrewAI, and AutoGen.
- Hybrid RAG Architecture – Combining dense vector search, sparse retrieval, keyword search, and reranking techniques.
- Model Context Protocol (MCP) – Building MCP servers and clients, tool integrations, and standardized AI application connectivity.
- LLM Fine-Tuning – Customizing foundation models for domain-specific applications.
- LoRA / QLoRA – Parameter-efficient fine-tuning and memory-efficient model adaptation.
- DPO / RLHF – Preference optimization and reinforcement learning from human feedback.
- Distributed GPU Inference – Multi-GPU deployments, parallel inference, batching, and high-throughput model serving.
- Prompt Engineering – System prompts, few-shot learning, prompt optimization, and context-window management.
3. Data, Vector Databases & Knowledge Management
- Pinecone – Vector indexing, similarity search, metadata filtering, and scalable retrieval systems.
- Weaviate – Vector database architecture, hybrid search, embeddings, and semantic retrieval.
- Qdrant / Milvus – Vector indexing, filtering, collections, and high-performance similarity search.
- PostgreSQL / pgvector – Vector storage, similarity search, relational data integration,
and hybrid retrieval.
- Neo4j Knowledge Graphs – Graph-based retrieval, entity relationships, and GraphRAG implementation.
- Redis Semantic Cache – Semantic response caching, session management, and low-latency retrieval.
- Apache Spark – Large-scale data processing, document transformation, and distributed data engineering.
- Delta Lake / Snowflake – Data ingestion, structured data processing, and enterprise data integration.
- Amazon S3 – Document storage, data lakes, file management, and AI knowledge-base integration.
4. Cloud, Model Deployment & Infrastructure
- AWS SageMaker – Model training, experimentation, deployment, and ML lifecycle management.
- AWS Bedrock – Foundation model access, model customization, knowledge bases, and managed AI applications.
- AWS EKS GPU Nodes – Kubernetes-based GPU workloads and scalable LLM deployments.
- Azure OpenAI – Enterprise LLM integrations, secure deployments, and production AI applications.
- GCP Vertex AI – Foundation model development, model training, deployment, and generative AI services.
- Docker – Containerization, application packaging, and deployment standardization.
- Kubernetes – LLM workload orchestration, GPU resource management, scaling, and high availability.
- NVIDIA Triton Inference Server – Production model serving, inference optimization, and concurrent model execution.
- vLLM – High-throughput LLM inference, continuous batching, and KV cache optimization.
- TensorRT-LLM – GPU-accelerated inference and model performance optimization.
- Terraform – Infrastructure as Code, cloud provisioning, and automated infrastructure management.
- GitHub Actions – CI/CD automation, model deployment pipelines, and application release management.
5. LLM Observability, Security & Evaluation
- LangSmith – LLM application tracing, debugging, monitoring, and execution analysis.
- Arize Phoenix – LLM observability, tracing, retrieval evaluation, and performance analysis.
- Ragas – RAG evaluation, context relevance, answer quality, and faithfulness measurement.
- DeepEval – LLM testing, evaluation metrics, and automated quality assurance.
- MLflow – Experiment tracking, model versioning, deployment, and lifecycle management.
- Weights & Biases – Model training experiments, metrics tracking, and performance visualization.
- NeMo Guardrails / Guardrails AI – LLM output validation, policy enforcement, and controlled AI interactions.
- Prompt Injection Defense – Detecting malicious prompts, jailbreak attempts, and unsafe model instructions.
- PII Redaction – Sensitive information detection, masking, and privacy protection.
- AI Security & Governance – HIPAA/SOC 2 considerations, access control, auditability, and responsible AI implementation.
6. Advanced Generative AI,
MLOps & Enterprise Integration
- GraphRAG – Combining knowledge graphs with retrieval-augmented generation for complex enterprise queries.
- Advanced Retrieval Optimization – Query rewriting, contextual retrieval, reranking, and metadata-based filtering.
- LLMOps – Automated model evaluation, deployment, monitoring, versioning, and lifecycle management.
- Model Quantization – Reducing model memory requirements using 4-bit, 8-bit, and other quantization techniques.
- Embedding Optimization – Selecting embedding models, vector similarity tuning, and retrieval performance optimization.
- AI API Integration – REST APIs, third-party services, enterprise systems, and external tool integrations.
- Distributed AI Systems – Scalable inference architectures, GPU scheduling, caching, and concurrency management.
- AI Cost Optimization – Token usage monitoring, model selection, inference optimization, and cloud GPU cost management.
- Enterprise AI Governance – Model access policies, audit logging, data privacy, and compliance controls.
- Technical Leadership – Architecture reviews, technical documentation, mentoring, and cross-team collaboration.
Candidate Requirements
- Minimum 6–9 years of professional experience in Software Engineering, Machine Learning, Generative AI, or AI Engineering.
- Strong hands-on production experience with Python, PyTorch, LangChain, and LlamaIndex.
- Proven experience developing and deploying production-grade LLM applications using OpenAI, Anthropic, or open-source foundation models.
- Strong understanding of RAG architectures, hybrid retrieval, embeddings, vector databases, and semantic search.
- Practical experience building Agentic AI workflows, multi-agent systems, and MCP integrations.
- Experience with LLM fine-tuning techniques such as LoRA, QLoRA, DPO, or RLHF.
- Strong knowledge of Vector Databases such as Pinecone, Weaviate, Qdrant, Milvus, or pgvector.
- Experience deploying AI applications using AWS, Azure, or GCP.
- Hands-on experience with Docker, Kubernetes, FastAPI, and production model-serving frameworks.
- Solid understanding of LLM evaluation, observability, prompt injection defense, and AI security.
- Experience optimizing LLM inference performance, GPU utilization, latency, and token costs.
- Ability to explain complex Generative AI architectures through real-world production project scenarios.
- Excellent English communication and technical documentation skills.
- Experience mentoring AI engineers and conducting technical interviews is an advantage.
- Based in India and available for approximately 10–15 hours per week of remote freelance/part-time consulting.
- Comfortable supporting multiple Generative AI, LLM Engineering, RAG, and AI Platform-related job positions within their areas of expertise.
Pay: ₹20,000.00 - ₹50,000.00 per month
Benefits:
- Flexible schedule
Experience:
- Python: 6 years (Preferred)
- PyTorch: 6 years (Preferred)
- LangChain: 6 years (Preferred)
- LlamaIndex: 6 years (Preferred)
- OpenAI API: 6 years (Preferred)
- Pinecone: 6 years (Preferred)
- FastAPI: 7 years (Preferred)
Work Location: Remote
📌 Senior Generative AI & LLM Engineer — Freelance / Part-Time Technical Support (India)
🏢 Abh Consultants
📍 India