Hey there! 👋
I'm Mohit Kumar — Senior Backend Engineer.
I architect resilient distributed financial systems. 👨💻
Specializing in event-driven architecture (Kafka), container orchestration (Kubernetes), and transactional safety. Currently building secure digital asset transfer workflows at Publicis Sapient / Galaxy Digital.
Let’s connect 👉 mohitkumar4706@gmail.com
I’m currently building backend systems at Publicis Sapient for Galaxy Digital, where I help move money, data, and crypto around safely and at scale. Think: Kubernetes migrations, Kafka pipelines, secure asset orchestration — the kind of stuff that breaks when you're not careful. I make it not break.
I’m obsessed with clean architecture, boring-but-beautiful APIs, and systems that scale without drama. I love working with early teams where nothing is figured out and duct tape still holds value.
When I’m not deep in logs or terminals, I’m probably lifting heavy things, riding fast bikes, or chasing one of too many side quests on my bucket list.
- Java
- Python
- SQL
- JavaScript
- TypeScript
- Spring Boot
- Spring Security
- Hibernate
- JUnit
- Mockito
- Maven
- Gradle
- MySQL
- PostgreSQL
- MongoDB
- Redis
- Elasticsearch
- DynamoDB
- Apache Kafka
- RabbitMQ
- AWS (EC2, S3, EKS, IAM, Lambda, CloudFront)
- Kubernetes
- Docker
- Jenkins
- Terraform
- CI/CD
- LLM Integration
- RAG
- Prompt Engineering
- NLP
- Scikit-learn
- LangChain
- Vector Databases
- Grafana
- Datadog
- OpenTelemetry
- SonarQube
- Microservices
- REST APIs
- System Design
- TDD
- Agile
- Built generalized crypto asset transfer tool for secure movement of digital assets between Fireblocks vaults, institutional exchanges, and counterparties, automating multi-million-dollar daily transactions that were previously manual.
- Developed threshold-based rebalancing engine that redistributes assets across accounts for active trading, lending, and operations, automatically derisking idle assets to cold vault storage.
- Building exchange connectivity layer that abstracts deposit/withdrawal address management from operations users via a unified dashboard, reducing operational burden and security risk.
- Led zero-downtime migration of 50+ microservices from EC2 to Kubernetes (EKS) with a three-phase dependency-aware rollout across production trading environments.
- Architected Kafka-based replay system for MSK to Confluent Cloud migration (25% cost reduction) and designed consumer-first bridge replacing RabbitMQ, eliminating 100% of pricing update failures.
- Architected backend for Apple Pay Later with HDFC Bank, building the eligibility engine, approval workflows, and real-time disbursement pipelines serving millions of users.
- Launched Flexi Credit lending platform processing INR 150M+ monthly across 85,000+ merchants with instant approvals up to INR 500K.
- Optimized merchant onboarding via Kafka, Redis, and multithreading, cutting processing time by 99.1% (hours to seconds).
- Built AML and fraud detection systems saving INR 360M+ in risk exposure; integrated HDFC, ICICI, Kotak APIs for real-time BNPL credit checks.
- Implemented CloudFront + Redis caching that cut P90 API latency 81% (800ms to 150ms) and reduced database load by 70% for a platform serving 10M+ monthly active users.
- Led Study IQ platform integration, designing unified data models and REST APIs for cross-platform revenue tracking.
Automated trading bot with LIVE/PAPER modes using ML-driven strategy selection and LLM-powered sentiment analysis via RAG pipeline over financial news. Implemented risk management with dynamic stop-loss, position sizing, and exposure limits; modular backtesting architecture reduced API costs 40%.
Designed and developed the architecture for Apple Pay Later, integrating multiple lenders like HDFC, ICICI, and Kotak to enable personalized financial options for lending platforms. Reduced merchant whitelisting time by 99.11% using Kafka, caching, multithreading, and algorithms.
Developed a common-consumption module to prevent site crashes under high load and implemented Redis for faster API response times, improving live class performance. Pioneered multilingual packages for better accessibility.
Engineered low-latency conversational bots for automated client contact centers, enabling seamless integration of service platforms like Call, WhatsApp, and Web chat.
Automated trading bot with LIVE/PAPER modes using ML-driven strategy selection and LLM-powered sentiment analysis via RAG pipeline over financial news, improving prediction accuracy by 20%. Risk management with dynamic stop-loss, position sizing, and exposure limits; modular backtesting architecture reduced API costs 40%.