Anshuman Pattnaik

Full-Stack AI Engineer with 3+ years of experience building production-grade AI, agentic and enterprise software systems across Python/FastAPI, React and cloud-native platforms.

Built

LangGraph-based multi-agent workflows combining RAG, tool calling, structured state and human-in-the-loop controls for operational and domain-intelligence applications. Strong engineering foundation across Kubernetes, microservices, machine learning, Kafka, Redis, MongoDB/PostgreSQL and secure API platforms.

Associate Software Engineer — AI & Platform Engineering at INTELLECT INFORMATICS (2024-05 – Present)

OpsBuddy – AI-enabled Enterprise Operations & Intelligence Platform

• Built an AI-enabled operations console spanning ~30 state deployments, integrating ticket intelligence, RAG, approvals, automation, and observability.
• Implemented hybrid RAG using Elasticsearch BM25, 768-dimensional kNN vectors, and RRF fusion for grounded retrieval across operational knowledge.
• Built heading-aware knowledge ingestion with ~800-token chunking, overlap, PII masking, metadata filtering, curator approval, and citation-enforced answers.
• Implemented grammar-first natural-language analytics with LLM fallback, controlled-vocabulary validation, and k-anonymity safeguards for sensitive banking dimensions.
• Built safe NL-to-SQL where LLM-proposed SELECT queries pass sqlglot classification, scope validation, and human approval before execution.
• Built ticket intelligence using pattern matching, recurrence detection, and controlled vocabulary informed by a 521,957-ticket support corpus.
• Automated SLA sweeps, ticket imports, knowledge ingestion, alert-to-ticket creation, notifications, and health checks using Celery workers and scheduled jobs.

Associate Software Engineer — AI & Platform Engineering at INTELLECT INFORMATICS (2024-05 – Present)

AI-enabled Intelligence Operator & Kubernetes Deployment Gateway

• Built a production-grade Kubernetes AI gateway for 28 deployed modules across an 11-VM cluster with 3 control-plane and 8 worker nodes.
• Implemented LangGraph-based multi-agent orchestration across triage, diagnosis, and action planning using shared structured state for incident workflows.
• Built conditional tool calling and context routing across logs, metrics, Kubernetes events, deployment history, and retrieved operational evidence.
• Built a RAG-grounded operational intelligence layer over pod logs, Kubernetes events, alerts, deployment metadata, runtime telemetry, and namespace signals.
• Designed human-in-the-loop approval controls before ticketing, notifications, rollback, remediation actions, or protected Kubernetes execution.
• Secured APIs with rate limiting, client credentials, and JWT; built React views for deployment, telemetry, and incident review.

Associate Software Engineer — AI & Platform Engineering at INTELLECT INFORMATICS (2024-05 – Present)

Farmer & PACS Portfolio Intelligence

• Built a production-grade agri-credit intelligence platform across 211 PACS and ~120,000 farmer records, analyzing loan utilization, repayment behavior, and portfolio movement.
• Engineered borrower, crop, loan, repayment, and overdue features from legacy PACS data for clustering, forecasting, trend analysis, and reporting.
• Applied K-Means clustering to segment farmers by repayment ratio, overdue frequency, loan exposure, crop seasonality, and land profile.
• Built season-aware analytics and XGBoost forecasting to track overdue trends, collection efficiency, repayment ratios, and portfolio movement across seasons.
• Implemented RAG with chunking, embeddings, and ChromaDB retrieval using policy context to generate grounded portfolio narratives for PACS and district stakeholders.
• Built React intelligence dashboards across farmer, village, PACS, and district views for trends, clusters, forecasts, repayment behavior, and policy narratives.

Associate Software Engineer — AI & Platform Engineering at INTELLECT INFORMATICS (2024-05 – Present)

Enterprise ERP Modernization

• Implemented service decomposition and API boundary cleanup while migrating selected monolithic ERP workflows toward independently deployable service-oriented modules.
• Integrated Keycloak-based authentication and role-aware access controls across protected ERP modules and internal service APIs.
• Implemented Redis caching for frequently accessed reference and configuration data, reducing repeated database reads and improving backend response consistency.

Trainee Consultant at INVENIO (2023-05 – 2024-04)

InTrack Modernization – Enterprise Microservices Platform

• Contributed to modernization of an employee operations platform by supporting Spring Boot microservices, Angular screens, backend service boundaries, and Level 2 DFDs for HR, travel, time tracking, and resource workflows.
• Implemented Azure AD-based SSO and Spring Cloud Gateway token validation filters to enforce request-level access control before routing traffic to downstream services.

Trainee Consultant at INVENIO (2023-05 – 2024-04)

UAE Pass Orchestrator – OAuth to SAML Identity Broker

• Contributed to a middleware identity broker connecting an OAuth provider with a SAML consumer for mobile authentication flows.
• Designed middleware APIs for token introspection, X.509 certificate handling, key generation, handshake workflows, and authentication exchange to support WSO2 replacement.

Bachelor of Technology in Computer Science and Engineering – GIET University, Odisha (2019 – 2023)