Machine Learning Engineer

AI / ML Engineer – SLM & RAG Specialist

Location: Trivandrum(Kerala)

Company: NeoITO

Experience: 5+ Years

About the Role

NeoITO is hiring an AI / ML Engineer to build and own an AI-powered Proposal & RFP generation system designed to transform meeting notes into structured, client-ready proposals within minutes.

You will be responsible for designing and managing the core AI layer, including the inference engine, RAG pipeline, embedding models, and compliance validation system.

Y

ou will collaborate closely with backend (Node.js) and frontend (React) engineers to deliver a production-ready AI system within a defined delivery timeline.

Key Responsibilities

Model Deployment & Inference

• Deploy and manage Small Language Models (SLMs) on on-premise GPU infrastructure.
• Configure and optimize LLM inference pipelines using frameworks such as vLLM or HuggingFace Transformers.
• Implement token streaming, continuous batching, and optimized sampling strategies for reliable text generation.
• Apply quantization techniques (GPTQ/AWQ) to reduce GPU memory footprint while maintaining inference performance.
• Monitor GPU health and performance metrics including VRAM usage, latency, and throughput

Retrieval-Augmented Generation (RAG)

• Design and implement RAG pipelines to enable context-aware proposal generation.
• Build text chunking pipelines and generate embeddings using sentence-transformer models.
• Store and retrieve vector embeddings using PostgreSQL with pgvector.
• Implement semantic similarity search to retrieve relevant historical proposal data.
• Continuously evaluate and optimize retrieval quality and performance.

AI-Driven Proposal Generation

• Design structured pipelines to generate multi-section proposals including:
• Executive Summary
• Project Scope
• Technical Approach
• Implementation Timeline
• Investment Summary
• Risk Mitigation
• Create section-specific prompts and templates for high-quality generation.
• Implement real-time streaming responses to backend services.
• Support partial regeneration of sections for iterative proposal refinement.

AI Quality, Validation & Compliance

• Develop a validation engine to ensure generated content meets compliance and quality standards.
• Implement rule-based checks including:
• Client name verification
• Budget reference validation
• Section completeness
• Sensitive data detection
• Support an optional AI-based review layer for deeper quality checks.
• Deliver structured feedback and annotations for use within editing workflows.

Prompt Engineering & Model Optimization

• Design and maintain structured prompts for classification, generation, and validation tasks.
• Conduct iterative prompt optimization to improve accuracy, tone, and consistency.
• Maintain prompt versioning and regression testing frameworks.
• Evaluate output quality through structured human evaluation metrics.

Fine-Tuning & Model Improvement

• Lead fine-tuning initiatives to improve model performance over time.
• Prepare and curate training datasets from finalized proposals.
• Implement LoRA / QLoRA fine-tuning strategies for efficient model updates.
• Track experiments and model versions using tools such as MLflow.

Collaboration & Engineering Practices

• Expose AI capabilities via FastAPI services consumed by backend applications.
• Collaborate with backend teams on job orchestration, queue processing, and event streaming.
• Implement unit tests and quality checks for ML pipelines.
• Contribute to containerized deployment environments using Docker.
• Support CI/CD pipelines with automated testing and linting workflows.

Required Skills & Experience

Large Language Models & AI Systems

• Hands-on experience with LLMs or SLMs
• Experience deploying models using vLLM, HuggingFace Transformers, or similar frameworks
• Knowledge of quantization techniques and inference optimization

RAG & Vector Search

• Experience building Retrieval-Augmented Generation pipelines
• Knowledge of vector databases such as pgvector, FAISS, or similar
• Familiarity with embedding models and semantic search

Programming & Frameworks

• Strong Python development experience
• Experience with FastAPI, Pydantic, and PyTorch
• Knowledge of libraries such as sentence-transformers, LangChain, or LlamaIndex

Infrastructure & GPU Systems

• Experience working with GPU-based model deployment
• Familiarity with CUDA environments and GPU monitoring
• Experience deploying applications with Docker on Linux environments

Databases & Storage

• Experience with PostgreSQL
• Familiarity with vector extensions or vector search databases
• Knowledge of object storage solutions such as S3 or MinIO

MLOps & Model Lifecycle

• Experience with LoRA / QLoRA fine-tuning
• Familiarity with experiment tracking tools
• Knowledge of dataset preparation and model evaluation

Nice to Have

• Experience working with Meta Llama models
• Familiarity with document generation systems
• Experience with queue-based ML pipelines
• Exposure to secure enterprise environments requiring strict data governance
• Knowledge of observability tools such as Prometheus

In this role, you will:

• Deliver a fully functional AI proposal generation system running entirely on-premise
• Achieve high-quality, structured proposal outputs
• Ensure stable performance under concurrent usage
• Establish a foundation for continuous model improvement through fine-tuning

Tech Stack

Primary Language: Python

API Framework: FastAPI

LLM Inference: vLLM / Transformers

Embedding Models: Sentence Transformers

Vector Database: PostgreSQL + pgvector

GPU Infrastructure: NVIDIA GPU environments

Containerization: Docker

Monitoring: Prometheus

Testing: Pytest