About Me
Software Developer building backend services, data pipelines, and on-device computing tools.
Pritam Das
Software DeveloperBased in Kolkata, India. Open to relocation & remote work. Specializing in backend APIs, SQL optimization, data pipelines, and developer tooling.
Over the past three years, I have engineered and maintained systems that organizations rely on daily — real-time bidding infrastructure at Agnik, an internal operations portal for insurance workflows, and production web applications. I specialize in backend engineering: designing predictable REST APIs, relational schemas with Knex migrations, and data pipelines that run reliably in production.
My approach to software development focuses on practical problem solving: taking repetitive, error-prone manual operations — such as reconciling advertising reports across multiple APIs, tracking daily work targets, or translating comic chapters — and replacing them with robust, automated software.
Currently exploring Rust · local AI runtimes
At Agnik, I work on advertising data pipelines and real-time bidding algorithms. My contributions include building a Real-Time Bidding (RTB) tool that automated campaign bid adjustments, tuning search ranking algorithms to improve product placement, building SQL data aggregation pipelines with Apache ECharts dashboards, and optimizing query performance across PostgreSQL and SQL Server.
In addition to my work at Agnik, I develop freelance client solutions and open-source software:
- Teammates Insurance: Full-stack insurance administration portal with Role-Based Access Control (RBAC) and asynchronous Redis Bull queue report generation.
- Gram Bahar: Direct-to-consumer eCommerce platform connecting 200+ rural jaggery farmers directly with consumers via dynamic combo checkouts.
- TabDent Pharma: High-speed pharmaceutical formulation catalog with active ingredient breakdowns and sub-150 KB page weight.
- Inori Studio: Browser-based image editor running background removal and inpainting on-device via WebGPU and WebAssembly.
- Work Tracker for Jibble: Manifest V3 browser extension for Chrome and Firefox syncing timesheet data and tracking daily work targets via toolbar badges.
- Comic Translator: Python CLI pipeline using OpenCV morphological dilation and PaddleOCR for automated speech bubble cleaning and typesetting.
Engineering Focus & Practical Priorities
Backend Reliability & Structured Data
Designing REST APIs, relational PostgreSQL schemas with Knex migrations, Redis Bull queues for asynchronous report generation, and granular Role-Based Access Control (RBAC).
On-Device AI & Client-Side Compute
Running model inference directly on user hardware with WebGPU compute shaders, WebAssembly Web Workers, and local LLM runtimes (Ollama / llama.cpp) to eliminate server inference costs and preserve data privacy.
Performance & Pipeline Optimization
Building data aggregation pipelines for advertising analytics, tuning search ranking logic, and optimizing SQL queries to reduce batch processing and dashboard turnaround time.
Developer Tooling & Automation
Building practical CLI tools, browser extensions (Manifest V3), and scrapers that automate repetitive workflows and streamline operational tasks.