Pranay Prasoon
Backend & ML Engineer
Backend-focused Software Engineer with strong expertise in distributed systems, data engineering, and applied machine learning. Experienced in building scalable backend architectures and production-grade ML pipelines.
About Me
I am a Backend-focused Software Engineer with strong expertise in distributed systems, data engineering, and applied machine learning. I have experience building scalable backend architectures, real-time systems, and production-grade ML pipelines, including NLP services and time-series forecasting models.
Adept at delivering end-to-end backend + ML solutions from data ingestion to deployment, I enjoy solving complex infrastructure challenges and optimizing performance for data-intensive applications.
Backend Systems
Architecting high-throughput, distributed services using FastAPI, Node.js, and Golang.
Data Engineering
Building pipelines with Kafka/Celery, and managing polyglot persistence (SQL, NoSQL, Time-series).
Applied ML
Deploying inference services, NLP pipelines, and forecasting models to production environments.
Featured Projects
Selected work conceptualizing and building complex systems.
PulseIQ – Real-Time Public Sentiment & Intelligence
- Architected a distributed data ingestion pipeline for real-time social/news data.
- Implemented async task orchestration (Celery/Redis) to decouple ingestion from ML inference.
- Built a dedicated NLP microservice with fine-tuned BERT models and spaCy NER.
- Designed polyglot persistence: MongoDB (docs), TimescaleDB (time-series), Elasticsearch (search).
- Containerized and orchestrated 6+ microservices using Docker Compose.
Crypto Price Prediction System
- Implemented and compared ARIMA and Facebook Prophet models for short/medium-term forecasting.
- Performed advanced data preprocessing: Stationarity checks, Trend/seasonality decomposition.
- Evaluated models using RMSE, MAE, and rolling-window validation.
- Designed for easy extension to real-time price ingestion and API-based predictions.
SafeChat – ML-Powered Secure Chat
- Developed a real-time chat platform with ML-powered message moderation.
- Implemented NLP models to detect toxic/unsafe messages and classify intent.
- Designed backend APIs to block, flag, or sanitize messages in real-time.
- Integrated ML inference into the chat pipeline with minimal latency impact.
Scalable Online IDE Backend
- Executed untrusted user code securely inside isolated Docker containers.
- Implemented real-time collaboration using WebSockets.
- Built project isolation, execution lifecycle management, and authentication.
- Designed for scalability to handle concurrent user sessions.
Technical Proficiency
A comprehensive toolkit for building end-to-end data-intensive applications.