Joybrata sarkar

SOFTWARE ENGINEER

ABOUT

aBOUT ME

I’m a full stack engineer who enjoys building practical, scalable systems — especially when AI is involved. Over the past few years, I’ve worked across the stack to deliver production-grade platforms in hiring and HR tech.

At Impacteers, I’ve been building Zooper Smart ATS, an AI-based recruitment platform. I worked on the full architecture using Docker-based microservices, built an interview module, optimized job description generation using Redis and LangGraph, and fine-tuned Wav2Vec2 for accurate speech transcription. My focus here has been on performance, real-time systems, and integrating ML models into production.

Before that, at Supersourcing, I helped ship 3 enterprise SaaS products — including an AI Interview system — while leading frontend development using Angular, NgRx, and TypeScript. I worked with FastAPI and LangChain on the backend to build real-time, low-latency AI pipelines for interviews, and I contributed to hiring and mentoring developers as the team grew.

I started my journey at BOT Shreyasi, where I worked on ATS systems and built tools like a visual campaign builder and resume parser integration.

Over time, I’ve moved from frontend-heavy work to full-stack and AI-focused systems. I’m comfortable across the stack, enjoy solving technical challenges, and like working on products that have real impact.

SKILLS

Areas of Expertise

Scalable Backend Systems: We design and implement scalable backend systems using cutting-edge technologies like FastAPI, Django, and Redis Streams, enabling real-time, event-driven applications that drive business growth and efficiency.
Intelligent NLP Systems: Our production-grade NLP pipelines, built with Hugging Face Transformers and LangChain, enable real-time question answering, summarization, and conversational agents, revolutionizing the way businesses interact with customers and process information.
Lightning-Fast Search: We engineer semantic and phonetic search pipelines using TF-IDF, cosine similarity, and Redis Search, ensuring low-latency, high-accuracy query resolution across structured and unstructured datasets, and transforming the way businesses retrieve and utilize data.
Real-Time Communication: Our real-time audio/video platforms, built with WebRTC and Redis-based streaming backends, enable seamless communication and collaboration, while integrating AI-powered audio intelligence for real-time transcription, voice activity detection, and more.
Microservices Architecture: We architect Dockerized microservices with environment-specific configurations and CI/CD pipelines, ensuring modular, maintainable, and scalable deployments that drive business agility, reduce costs, and improve overall system reliability.

December 2024 – Present

Full stack engineer / Impacteers

Job Description Creation: We engineered a game-changing Job Description Creator using LangGraph and adaptive RAG, slashing generation time from 4 seconds to 0.1 milliseconds. This innovation enables recruiters to work more efficiently and effectively.
Architected and deployed a scalable microservices-based system for Zooper Smart ATS Designed the project architecture using Creational Patterns, ensuring modularity, robust error handling, comprehensive logging, secure authentication, effective rate limiting, and environment-based configurations
AI-Powered Candidate Scoring: We developed a cutting-edge candidate scoring and reporting system that combines lexical, phonetic, and semantic search to compute candidate-job relevance scores. This system provides recruiters with clear, data-driven insights, reducing evaluation time and improving hiring accuracy.
Real-Time AI Interviewer Platform: Our team architected an AI Interviewer Platform that leverages WebSocket, Redis Streams, Celery, and FastAPI for real-time session handling and bi-directional communication. This platform enables instant feedback and seamless interaction between candidates and recruiters, enhancing the overall hiring experience.
Lightning-Fast Data Retrieval: We implemented MongoDB vector storage and Redis Search with key-based and semantic caching techniques, enabling fast and accurate semantic retrieval. This innovation ensures ultra-low latency and instant responses, making it easier for recruiters to access the information they need to make informed decisions.

June 2022 – December 2022-November 2024

Frontend Engineer / Supersourcing

Scalable Solutions: Successfully shipped three SaaS products, including an AI interview platform, ATS, and vendor management tool, ensuring they can handle growth and high traffic without compromising performance.
Performance Boost: Optimized front-end architecture using NgRx state management, lazy loading, and component-level caching, resulting in enhanced performance even under heavy user load - a game-changer for user experience.
AI-Powered Innovation: Engineered an interactive AI interview platform with cutting-edge features like dynamic question generation, real-time sentiment analysis, and concurrent multi-candidate support, making hiring processes more efficient and effective.
Real-Time Excellence: Designed a Redis-based session management system that improves state tracking and reduces latency in real-time interactions, ensuring seamless and responsive user experiences.
Code Quality Matters: Reviewed and optimized the codebase to ensure best practices in maintainability, readability, and efficiency, setting the foundation for future innovation and growth - a testament to the importance of clean, high-quality code.