Nikhil Sai Nambula
I build Java and Spring Boot services that stay up when a dependency does not. Three years in production on a platform serving 2 million users.
- in production
- users served
- requests a day
- LeetCode solved
91.4%
-40% primary DB load40 → 8
N+1 fetch eliminationEngineering Mindset & Production Philosophy
Systems engineering grounded in measurement, resilient failure paths, and readable architectures that scale cleanly under production pressure.
I am a Java backend developer at Tata Consultancy Services, where I own three Spring Boot microservices end to end on AWS. The platform serves over 2 million users and about 50,000 requests a day.
Most of my work sits in the gap between working and keeping working. I redesigned JPA fetch strategies to kill an N+1 query pattern, which took a request from roughly 40 database round-trips down to 8. I built a Redis caching layer with per-key TTL policies that cut primary database load by 40%. I put circuit breakers, retry with exponential backoff and dead-letter queues in front of services that are allowed to fail.
Outside of work I grind DSA and design systems, which is where the 1000+ LeetCode problems and the low-level-design playground come from. AWS Certified Developer, Associate, 2026.
Java Backend Developer · TCS
Non-Negotiable Architecture Principles
Measure first
Every performance number I quote came off a query plan, a profiler or a load test.
Design the failure path
Timeouts, retries, poison messages and partial writes are where systems actually live.
Boring technology
A well-understood Postgres beats a distributed novelty at 3am.
Readable over clever
SOLID and named patterns exist so the next person can extend the code.
Technical Arsenal & Core Competencies
Applied in real production systems serving millions of requests, focused on backend resilience, clean architecture, and distributed data pipelines.
Spring Boot
Primary enterprise framework powering 3 microservices with Resilience4j circuit breakers, Spring Security, and REST APIs.
Languages
Low-level design
Frameworks & APIs
Distributed Architecture
Data stores
AWS Infrastructure
DevOps & Testing
Distributed Systems & Reliability Lab
Interactive models demonstrating resilience patterns, caching strategies, and event-driven architecture designed for high-availability production.
Downstream Dependency Fault Tolerance
Adjust simulated failure rates to observe the state transition between normal execution, fail-fast circuit opening, and canary probing.
Downstream failure rate is under the 50% threshold. Requests are passed directly to database and external services with full throughput.
16 ms
Normal processing99.98%
Zero unhandled 500sDirect DB Connection
Synchronous@CircuitBreaker(name = "orderService", fallback = "fallbackOrder")
@Retry(name = "orderRetry", maxAttempts = 3)
public OrderResult process(OrderReq req) { ... }
Java Backend Engineering & Microservices
Three years building and maintaining high-throughput Spring Boot services on AWS, serving over 2 million users and 50,000+ daily requests.
Java Backend Developer
Tata Consultancy Services (TCS)·Hyderabad, India
Core Platform Mandate
Three Spring Boot microservices owned end to end, from design through deployment and monitoring. Serving 2 million active users and 50,000+ requests daily.
Checkout endpoints suffered from excessive database roundtrips due to default Hibernate EAGER entity relations. Redesigned query strategies using batching and JPQL JOIN FETCH to preserve database health under peak burst traffic.
- • ~40 database roundtrips / req
- • N+1 query loop on items
- • P95 Latency: 165ms
- • DB Connection pool exhaustion
- • 8 database roundtrips / req
- • LAZY + @BatchSize(25)
- • P95 Latency: 24ms (-30%)
- • 40% reduction in DB pool wait
Production Architecture Achievements
- Reduced API latency by 30% by redesigning JPA and Hibernate fetch strategies, cutting database round-trips per request from about 40 to 8.
- Cut primary database load by 40% with a distributed Redis caching layer using custom per-key TTL policies, which also improved P99 throughput under burst traffic.
- Decoupled three downstream services by moving them onto a Kafka event-driven architecture instead of synchronous calls.
- Implemented JWT and role-based access control across the REST APIs, and wrote the JUnit 5 and Mockito suites for the services I own.
Engineering Projects & System Designs
Systems built from first principles focusing on concurrency, high availability, fault-tolerance, and practical architectural patterns.
ShopVerse
Distributed e-commerce backend platform
A distributed commerce platform built from scratch. Orders move through an explicit finite state machine rather than mutable status flags, and sessions live in Redis cluster with token-bucket rate limiting.
- Java, Spring Boot and MongoDB on the backend, structured strictly around SOLID and clean architecture.
- Stateless JWT authentication paired with Redis session management and automatic token revocation.
- Optimistic locking mechanism on inventory models preventing duplicate deductions during high concurrency.
- Automated CI/CD deployment pipeline with GitHub Actions, containerized with Docker.
Recipe generation from food photos
Computer vision & deep learning pipeline
An end-to-end vision pipeline that parses raw dish images and predicts ingredient compositions. InceptionV3 and Xception convolutional feature encoders feed an LSTM sequential decoder.
- Achieved 87% top-5 accuracy on a custom-annotated subset of the Food-101 benchmark.
- Implemented transfer learning with fine-tuned top layers over dual frozen convolutional backbones.
- Sequence decoding optimized using beam search with a width of 3 to output coherent ingredient lists.
ApplyPilot
AI Job Application Autofill Chrome Extension (Manifest V3)
A deterministic autofill Chrome extension (Manifest V3) with semantic field ontology and dynamic verification. Automatically identifies and accurately fills multi-page job applications using an intermediate representation (FormIR) and per-kind filling strategies without blind auto-submitting.
- Built as a modular monorepo with Manifest V3 (background worker, content scripts, React side panel) paired with a Fastify API backend.
- Engineered a deterministic 5-step pipeline: FormIR page scanning, semantic FieldMatcher ontology lookup, and per-kind AutofillExecutor.
- Validated against a golden-question benchmark suite (126 entries across 35 intents) achieving >=95% accuracy, <5ms latency, and zero false positives.
- End-to-end browser test automation with Playwright against a synthetic 7-page ATS simulator handling CAPTCHA, OTP, and dynamic DOM mutations.
Credentials, Problem Solving & Education
Strong theoretical fundamentals in algorithms and distributed computing backed by recognized certifications and continuous problem solving.
LeetCode: 1,000+ Algorithmic Problems Solved
1042+
Across all categories656
63% advanced difficultyTop 3.8%
Top competitive percentile1860+
Biweekly & Weekly Contests- • 0/1 & Unbounded Knapsack
- • Longest Common Subsequence (LCS)
- • Matrix Exponentiation & State DP
- • Interval & Tree DP
- • Coin Change & Knapsack variations
- • Edit Distance & Regular Expression Matching
- • Burst Balloons (Hard)
SRKR Engineering College
B.Tech, Computer Science and Engineering
Bhimavaram, Andhra Pradesh · GPA 8.0 / 10.0
Key Computer Science Coursework
- Data structures and algorithms
- Operating systems
- DBMS & Transaction Management
- Computer networks
- Object-oriented analysis and design
- Distributed computing
- Cloud computing
- Design and analysis of algorithms
AWS Certified Developer, Associate
Core cloud mastery covering EC2, Lambda serverless, DynamoDB, API Gateway, SQS queues, and IAM least-privilege security.
1000+ LeetCode Solved
Top 3.8% globally. Deep focus on dynamic programming, topological sorting, graph cycle detection, and memory-efficient algorithms.
Direct Engineering Channel & Collaboration
Actively open to full-time backend and distributed systems engineering roles. Send a direct message or connect across professional channels.
Send a Direct Message
Opens your email app with the message pre-filled and addressed to Nikhil.
Direct Channels
Preferred for immediate contact or recruiter outreach.
Hyderabad, India (IST / UTC+5:30)