SM

Hello

Suman Mohanty

Senior Software Engineer · Bengaluru, IN

I turn retrieval-heavy AI ideas into backend systems that can be audited, operated, and trusted in production.

download resume get in touch
5+
years
6
selected systems
40
stack items
System map RUNNING

Context path

GraphRAG gathers evidence before the model reasons.

retrieval.layer

GraphRAG Retrieval

Pulls operational context from runbooks, incidents, dependencies, and semantic graph data before the agent reasons.

Composed of GraphRAG · Vector index · Re-ranker · Citation builder

About

AI/ML engineer building production LLM systems — agents, GraphRAG, and neuro-symbolic pipelines that combine probabilistic inference with deterministic rules. Over five years I've shipped licensing infrastructure for millions of users, AI code-review platforms, customer APIs, and graph-grounded reasoning systems. My strongest work sits at the point where product pressure meets reliability: explainability, latency, observability, and the engineering that keeps agents trustworthy.

Experience

Senior Software Engineer, Data and AI

NuStudio.ai

Mar 2026 — Present · Bengaluru, IN
  • Established semantic data modelling and graph reasoning standards, integrating neuro-symbolic validation layers that separate probabilistic LLM inference from deterministic business rule enforcement — improving explainability, auditability, and production safety.
  • Designed and built an AI agent system using GraphRAG, Neo4j, and multi-agent orchestration to enable contextual retrieval and reasoning across operational data (services, dependencies, incidents, runbooks). Implemented long-term memory, vector search, async FastAPI pipelines, and tool orchestration focused on scalable and reliable agent execution.

Senior Software Engineer

Tally Solutions

Nov 2023 — Mar 2026 · Bengaluru, IN
  • Modernized the licensing & inference ecosystem by decoupling a legacy monolith into FastAPI, Node.js, and gRPC microservices powering LLM-based decision systems. Integrated Neo4j knowledge graphs for entity linking and reasoning over enterprise data. Migrated workloads to AWS Lambda/ECS Fargate, contributing to an estimated $1.8M in annual infrastructure savings (internal estimate) with Docker blue-green deploys.
  • Delivered an AI-powered code review & security platform integrating LLM-based static analysis and multi-scanner detection (OSV, OWASP); automated Jira ticketing and Bitbucket PR checks, reducing manual review time by 70% with Vertex AI for contextual insights.
  • Built real-time inference APIs for GenAI, ML, and custom models powering predictive maintenance, fraud detection, and RAG workflows. Conducted model training and fine-tuning; analyzed customer call recordings for sentiment and resolution scoring, optimizing latency with caching and parallel inference.
  • Led Customer Portal API development for 2.5M+ users using Django — owning LLDs, ETL pipelines, and real-time sync between PostgreSQL and Salesforce via Kafka, Debezium, and CDC.

Software Development Engineer

Boston Scientific (ITC Infotech)

Jan 2021 — Jun 2023 · Gurugram, IN
  • Simplified Windchill PLM tasks, contributing to a program valued internally at ~$3.3M, using Java, Spring Microservices, and AWS Serverless; optimized DB schemas and SQL query tuning.
  • Reduced support tickets by 65% with the RNA Bot (Python + Rasa chatbot). Cut Windchill licensing costs by 45% via a serverless document-sharing utility on AWS Lambda with Azure AD SSO.

Additional Experience

Software Engineer L2

Relevance Labs

Engineered scalable gRPC microservices for Polysign Crypto Wallet with BDD test automation, improving maintainability and development velocity.

Jul — Oct 2023 · Bengaluru, IN

Selected Work

Recent projects across NuStudio.ai and Tally — production AI systems, knowledge-graph reasoning, and the platform engineering that keeps them reliable.

NuStudio.ai

GraphRAG Multi-Agent System

AI agent system that reasons over operational data — services, dependencies, incidents, runbooks — using GraphRAG retrieval and multi-agent orchestration. Long-term memory, vector search, and tool orchestration built on async FastAPI.

↳ Contextual retrieval and reasoning across operational graph data

GraphRAGNeo4jFastAPIVector SearchMulti-Agent

NuStudio.ai

Neuro-Symbolic Validation Layer

Validation framework separating probabilistic LLM inference from deterministic business rule enforcement. Sets semantic data modelling and graph reasoning standards for production AI.

↳ Improved explainability, auditability, and production safety

LLMKnowledge GraphsSymbolic Reasoning

Tally Solutions

AI Code Review & Security Platform

LLM-based static analysis combined with multi-scanner detection (OSV, OWASP). Automated Jira ticketing and Bitbucket PR checks with Vertex AI for contextual insights.

↳ 70% reduction in manual code review time

LLMVertex AIOSVOWASPBitbucketJira

Tally Solutions

Licensing & Inference Microservices

Decoupled a legacy monolith into FastAPI, Node.js, and gRPC microservices powering LLM-based decision systems. Integrated Neo4j knowledge graphs for entity linking; migrated to AWS Lambda/ECS Fargate.

↳ $1.8M annual cost savings

FastAPIgRPCNeo4jAWS LambdaECS FargateDocker

Tally Solutions

Customer Portal API

Django-based API serving 2.5M+ users, owning LLDs, ETL pipelines, and real-time sync between PostgreSQL and Salesforce via Kafka, Debezium, and Change Data Capture.

DjangoPostgreSQLSalesforceKafkaDebeziumCDC

Boston Scientific

RNA Bot — AI Support Chatbot

Python + Rasa chatbot deployed inside Windchill PLM workflows. Paired with a serverless document-sharing utility on AWS Lambda + Azure AD SSO.

↳ 65% reduction in support tickets; 45% lower licensing costs

PythonRasaAWS LambdaAzure AD SSO

Skills & Certifications

AI / ML

Hugging Face LangChain LangGraph Langfuse Ollama GPT Vertex AI OpenCV TensorFlow Scikit-learn NumPy Pandas

Skills

LLM Training & Inference, Transformers, GraphRAG, RAG, Embedding Retrieval, GenAI Pipelines, Fine-Tuning

Backend

FastAPI Django Node.js Spring gRPC

Programming

Python Java C++

Data

Neo4j PostgreSQL Kafka MySQL MongoDB Debezium SQL CDC

Cloud / DevOps

AWS Lambda ECS Fargate Docker Serverless Framework CI/CD

Achievements & Certifications

GCP ML Engineer Professional · RASA · DeepLearning.AI · IBM Machine Learning · IoT — Yonsei University · KPMG Data Analytics · DSA — LinkedIn Learning

Education

2017 — 2021

Sikkim Manipal Institute of Technology

B.Tech, Computer Science Engineering

Sikkim, IN · 7.5 CGPA

Get in Touch

Best fit: senior backend or AI platform roles where LLM systems need retrieval, graph context, observability, and production guardrails. Email is the fastest route; LinkedIn works well for recruiter or hiring-team context.

EMAIL mohantysumansm@gmail.com LINKEDIN linkedin.com/in/sumanmohanty07 ↗ GITHUB github.com/sumanmohanty ↗