SDE · Broadridge · Hyderabad, IN Open to SDE-2 Roles

Mayank Sahu

> Java · Spring Boot · Distributed Systems · AI Infrastructure

Designing systems that scale, APIs that hold, and agents that remember.

3+ years engineering production backend services at Broadridge — microservices, event-driven pipelines, and on-call reliability for 10,000+ enterprise users. Now extending into AI infrastructure: agent memory, RAG pipelines, and multi-agent orchestration.

Resume PDF Email
10k+ Enterprise Users Served
~45% API Latency Reduced
3x Service Scalability
99.5% CI/CD Pipeline Uptime
50% MTTD Reduction

Featured System

Continuum

Active · v2.0.3 · MCP server
Memory infrastructure for AI agents — now an MCP server

Python · Postgres · pgvector · MCP · HNSW · BM25 · RRF · bge-m3

Agent memory is a hard problem — LLMs have no native persistence across sessions. Continuum solves it with a layered STM / MTM / LTM architecture over Postgres + pgvector, and ships an MCP server that gives any client — Claude Code, Claude Desktop, Cursor — four memory tools: remember, recall, current, timeline.

  • MCP server: Memory as tool calls over stdio or Streamable-HTTP — pip install continuum-mcp, register once, and the agent remembers. Published on PyPI.
  • Hybrid retrieval: Dense bge-m3 vectors + BM25 lexical search fused with Reciprocal Rank Fusion, over an HNSW-indexed pgvector store — matches on meaning and on exact tokens.
  • Bi-temporal & verified: Live-row invalidation with "as of" time-travel queries; 100% supersession and bi-temporal correctness, ~74% on LongMemEval-S, 1800+ tests green in CI.
Deep Dive ↗ Continuum MCP ↗ GitHub Repository →
memory.architecture
STM Session · Hot Context
↓ promote
MTM In-Process · Consolidation
↓ consolidate
LTM Postgres + pgvector · Hybrid Search

Experience

Jun 2023 — Present Hyderabad, IN

Software Engineer

Broadridge India

Platform · Reliability · Productivity

  • Engineered backend microservices in Java + Spring Boot backed by MySQL, serving 10,000+ enterprise users across financial workflows — reliability and scalability at the core.
  • Optimized REST API performance and inter-service communication, reducing aggregate latency by ~45% and improving service scalability by 3x.
  • Integrated Kafka and async messaging for high-throughput service-to-service flows, enabling event-driven patterns across distributed backend services.
  • Production on-call rotation on AWS — debugged across services, queues, and infra; reduced Mean Time To Detect (MTTD) by 50%.
  • Designed and operated Jenkins + Docker CI/CD pipelines for 15+ microservices at 99.5% uptime; Python automation frameworks eliminated 60% of manual ops.
  • Drove regression coverage to 85% via JUnit / PyTest, cutting release cycle time from 2 days to 4 hours.
JavaSpring BootMySQLKafkaAWSDockerJenkinsPythonREST APIs
Jan 2022 — Sep 2022 Bhubaneswar, IN

Product Engineering Intern

HighRadius

Full-Stack · AI · Fintech

  • Built full-stack features (React + Node.js + MongoDB) for 500+ corporate clients across 3 Agile releases.
  • Developed AI invoice-matching model (Python + scikit-learn), reducing manual processing by 80%.
ReactNode.jsMongoDBPythonML

Technical Stack

{ }

Languages

Java Python SQL JavaScript
⚙

Backend & Systems

Spring Boot REST APIs Microservices System Design Design Patterns Distributed Systems
⬡

Databases & Caching

MySQL MongoDB SQL Server Redis Vector DBs
⇄

Messaging & Events

Kafka Message Queues Event-Driven Async Patterns
☁

Cloud & DevOps

AWS Docker Jenkins CI/CD Observability Monitoring
◈

AI Infrastructure

OpenAI API Embeddings RAG Vector Search LangChain Multi-Agent LLM Orchestration
◉

Reliability

On-Call Incident Response MTTD/MTTR Production Debugging
✓

Testing & Quality

JUnit PyTest Mockito API Testing SonarQube Code Review

Education

Siksha 'O' Anusandhan University

B.Tech · Computer Science & Engineering · CGPA 8.66 / 10

2019 — 2023