Description

Job Description:-

Qualifications & Experience

• 5+ years in software development; 2+ years on AI/ML, NLP, or document processing solutions.
• Proven experience building RAG pipelines or LLM-integrated applications in Python.
• Hands-on experience with vector databases in production or PoC context.
• Domain experience in regulated industries (healthcare, insurance, financial services) is a strong plus.
• Strong debugging skills across OCR, embeddings, LLM responses, and application logic.
• Able to write clear deployment guides and API specifications.

Required Technical Skills

• Python (advanced): FastAPI/Flask for API development, pandas/numpy for data processing, LangChain or LlamaIndex for RAG orchestration.
• LLM APIs: Azure OpenAI (GPT-4o/GPT-4), Anthropic Claude, or equivalent — text classification, embedding generation, prompt engineering for multi-label tasks.
• Vector databases: OpenSearch, Pinecone, Weaviate, or Chroma — index creation, mapping configuration, k-NN vector search, bulk ingestion, query optimization.
• OCR/document processing: Azure AI Document Intelligence, Tesseract, or ABBYY for scanned and handwritten text. Pre-processing techniques (deskew, noise removal, layout detection).
• Low-code app platforms: Power Apps, Palantir Foundry, or Retool — build functional HITL review interfaces with forms, data display, and action buttons.
• RPA platforms: UiPath Studio for ingestion automations and enterprise system integrations. UiPath Orchestrator for job scheduling and queue management.
• Git and CI/CD: version control, CI/CD pipelines via Azure DevOps or UiPath Automation Ops for automated build, test, and deploy.
• REST API development: endpoints for document ingestion, classification retrieval, and review submissions.
• Enterprise integration: SharePoint, Confluence, content management systems — read/write connectivity.
• SQL: structured data stores for audit logging, review decisions, and classification result storage.

Key Responsibilities

• Build document processing pipelines: implement OCR extraction for handwritten and scanned documents, text normalization, cleaning, and structured output generation.
• Develop the vector database layer: create index schemas, compute embeddings using LLM embedding models, store documents with metadata, and optimize retrieval performance.
• Implement AI classification engines in Python: retrieve domain knowledge from vector stores, inject as context into LLM prompts, execute multi-label classification with confidence scoring.
• Build API ingestion layers: develop REST APIs and/or RPA integrations to ingest documents from enterprise content management systems and collaboration platforms.
• Develop human-in-the-loop validation interfaces: display AI classification results and confidence scores, enable SME accept/override, and log all decisions for audit.
• Tune confidence score thresholds to balance auto-classification vs. human routing for optimal accuracy and throughput.
• Write unit, integration, and end-to-end tests across all pipeline components: OCR accuracy, embedding quality, classification precision, and HITL workflow.
• Configure Dev and QA environments: set up OCR engines, vector databases, LLM connections, and required application access.
• Implement logging, monitoring, and error handling across the full pipeline for production reliability.
• Maintain technical documentation: code docs, API specifications, deployment runbooks, and configuration guides.
• Participate in iterative reviews and scope alignment with client stakeholders during PoC execution.
• Collaborate with business analysts to validate that classification logic aligns with domain-specific business rules and policies

Key Skills

Education

Any Graduate