HYDERABAD, INDIA
I build production agents with LangGraph, LLMs, and Phoenix.
Autonomous Sprint & Delivery Operations Agent
Reduced weekly sprint checks and ticket hygiene from 5 hours to 30 mins via LangGraph, Claude, and 50+ MCP tools.
Intelligent Workforce Helpdesk & Support Agent
Conversational Teams agent querying 2,000+ ServiceNow articles, in-chat catalog submissions, and role-based views for ICs and Managers.
High-Throughput Event Streaming & Analytics Platform
High-scale Kafka event streaming & DynamoDB caching, with full span-level tracing in Arize Phoenix & OpenTelemetry.
01 Timeline
August 2025 - Present
Align Technology · Hyderabad, India
April 2024 - July 2025
Locobuzz Solutions Pvt. Ltd. · Mumbai, India
02 Toolkit
Architecting autonomous multi-agent loops and state machines with LangGraph and MCP, connecting LLMs to enterprise tools like ServiceNow and Microsoft Graph.
Core Technologies & Tools
Deploying production LLM workflows using Claude, Azure OpenAI, and open models, specializing in structured schema outputs, prompt optimization, and task-specific fine-tuning.
Core Technologies & Tools
Engineering low-latency retrieval systems with dense vector search, semantic embeddings, custom chunking algorithms, and CDC-based knowledge synchronization.
Core Technologies & Tools
Instrumenting agentic and LLM applications with Arize Phoenix and OpenTelemetry for distributed trace inspection, latency monitoring, and production debugging.
Core Technologies & Tools
Building high-throughput asynchronous Python microservices with FastAPI, scalable Apache Kafka event streams, and Airflow orchestration for resilient data processing.
Core Technologies & Tools
Containerizing and deploying resilient cloud workloads across Azure and AWS with Docker, automated CI/CD pipelines, and serverless compute.
Core Technologies & Tools
Designing scalable data layers combining distributed relational stores, analytical OLAP databases, and multi-tier DynamoDB and Redis caching.
Core Technologies & Tools
Writing clean, maintainable Python and SQL for heavy data transformation, automated schema validation, and statistical analysis.
Core Technologies & Tools
03 PORTFOLIO
Sprint checks, ticket hygiene, leave tracking, and stakeholder reports were done by hand across ServiceNow, Jira, and Microsoft Graph.
A LangGraph agent dynamically routing between Claude Sonnet and Opus based on query complexity, with 50+ MCP tools running sprint and hygiene workflows traced in Phoenix.
5 hrs → 30 mins
Weekly sprint work that took about 5 hours now takes about 30 minutes (90% manual effort saved).
Business users needed to batch-migrate legacy enterprise articles (PDF, ASPX, DOCX) from SharePoint into ServiceNow KB without manual copy-pasting.
Automated ServiceNow catalog ingestion parsing Excel rows 1-by-1, executing a unified Azure Function for multi-format extraction (DOCX PDF export, PDF text/image parsing, ASPX asset separation), utilizing Azure AI Foundry LLMs for semantic HTML conversion of both PDF and ASPX articles.
2,000+ Human Hrs Saved
Saved 2,000+ human hours of converting PDF & ASPX text to HTML and creating articles in ServiceNow.
Employees spent excessive time searching through 2,000+ fragmented KB articles, navigating complex catalog ordering, and tracking ticket updates across portal silos.
An AI agent in MS Teams offering conversational Q&A over 2,000+ articles, in-chat catalog item requests, real-time ticket push notifications, and adaptive role-based views for Individual Contributors and Managers.
50% Faster Resolution
Cut query resolution time by 50% with instant conversational answers and proactive ticket status updates directly in Teams.
Referrals needed resume reading, skill sorting, acknowledgements, and status tracking by hand.
Microsoft Forms and SharePoint feed an LLM that parses resumes, classifies skills, sends acknowledgements, and tracks status.
Form to tracked status
A referral moves from the form to a tracked status without a manual handoff.
Batch processing could not keep up with thousands of concurrent users.
A Kafka pipeline handles ingestion, transformation, and failures. Airflow runs scheduled and on-demand jobs.
80% less processing time
Processing time dropped by 80%. Reporting latency dropped by 80%.
04 Education
SRM Institute of Science and Technology
Sept 2020 - May 2024 · GPA 9.2 / 10.0
05 Offer
Production-ready agents and multi-agent workflows for real business processes.
MCP servers connecting agents to enterprise APIs, databases and business tools.
RAG pipelines, semantic search, embeddings, knowledge ingestion and context-aware apps.
Python/FastAPI microservices, REST APIs, async and event-driven systems.
Document processing, knowledge management, enterprise workflows and repetitive ops.
Reliable agentic systems with clear boundaries, tool orchestration and failure handling.
Phoenix tracing for agent runs, tool calls, and LLM spans, plus testing and deployment.
06 Contact
Open to AI and backend roles, and to agent projects.
Phone
9979849768Location
Hyderabad, India