Engineering intelligent systems
for an autonomous future.
Zero-Trust Security Control Plane
for Autonomous AI Agents
Autonomous agents can invoke APIs, access infrastructure, use credentials, and execute real-world actions.
Traditional application security was not designed around autonomous machine identities making dynamic decisions.
AgentShield places a zero-trust control plane between an AI agent and the systems it attempts to access.
Every agent request is verified, evaluated, authorized, and recorded before execution.
Identity-aware agent sessions
OPA / Rego policy enforcement
Risk evaluation before execution
Kafka-backed immutable audit events
Human approval for sensitive actions
Agent containment and session revocation
A production commerce platform connecting storefront, payments, inventory, fulfillment, and intelligent wholesale workflows.
Payment, inventory, fulfillment, and customer communication move through one coordinated transaction lifecycle.
A live storefront backed by payments, inventory, fulfillment, event-driven workflows, and customer communication.
Payment providers can retry events. Customers can refresh, reopen tabs, or submit again. Fulfillment must stay consistent.
An AI-assisted clinical reasoning platform combining LLM-based reasoning, structured agent workflows, and PyTorch-powered classification.
Clinical information moves through an orchestrated reasoning pipeline before a structured response is produced.
Structured clinical features are passed through a PyTorch classification layer to produce a triage prediction with confidence scores.
Model output is evaluated against confidence thresholds and safety rules before it can become a clinical decision-support response.
A backend-first AI clinical assistant combining Claude reasoning, PyTorch triage classification, MCP-based agent tooling, structured logging, and auditable monitoring.
An AI-powered e-learning platform using OpenAI-based content recommendations to personalize learner experiences at scale.
Learner activity moves through the Django REST platform and OpenAI recommendation pipeline to deliver personalized content.
A scalable AI e-learning platform combining Django REST, React, MySQL, and OpenAI-powered content recommendations.
I'm Dhananjay Patel, an AI/ML and agent engineer building LLM-powered applications, agentic systems, backend infrastructure, and production AI platforms.
I enjoy building systems that connect machine intelligence with real-world engineering from AI agents and machine learning pipelines to secure distributed platforms, cloud infrastructure, and production applications.
Professional experience spanning backend engineering, cloud-native systems, distributed infrastructure, AI/ML integration, and production software.
Engineered Go microservices and cloud services across GCP, AWS, and Azure, with containerized deployments, event-driven Kafka pipelines, PyTorch integration, CI/CD automation, testing, and production issue analysis.
Developed backend APIs and integrated them with React interfaces, contributing to REST-based application development, Docker and Kubernetes environments, testing, and Agile software delivery.
AI engineering, backend systems, cloud infrastructure, distributed platforms, data pipelines, and production reliability.
Interested in AI/ML engineering, agentic systems, backend infrastructure, distributed systems, and production AI platforms.