Hi 👋 I’m Avishek Devnath, a backend-focused Software Engineer with hands-on experience designing and owning systems under real operational constraints. I currently work as a Senior CS Instructor at Phitron, but my primary identity is as a system-building engineer, not an educator.

I specialize in taking systems from unclear requirements to production architectures, owning design decisions around data flow, authorization, scalability, and long-term maintainability, and living with those decisions as systems evolve.

I’ve designed and built internal and public-facing platforms used in active environments, owning systems across their full lifecycle:

• architecture and service boundaries
• backend APIs and data modeling
• authentication and RBAC-heavy authorization flows
• background processing and long-running tasks
• iteration, maintenance, and refactoring after launch

These systems were deployed, evolved, and maintained over time, shaped by real user behavior and increasing complexity rather than static specifications.

SMS-Web — Role-Based Student Management Platform
Designed and built a backend-led, RBAC-heavy system for managing students, invitations, and administrative workflows. Owned data modeling, authentication, authorization logic, and API design, with a focus on correctness, access control, and long-term maintainability as requirements evolved.

BrainScale CRM — Internal CRM & Workflow Platform
Designed backend systems for an internal CRM platform supporting structured workflows, user roles, and operational data tracking. Focused on service boundaries, permission models, and evolving data schemas, ensuring the system remained adaptable as business needs changed.

These platforms represent core product work where reliability, permissions, and data integrity mattered more than novelty.

Designed an offline, consent-based biometric face recognition system, where offline inference was deliberately chosen to eliminate data exfiltration risk and simplify trust boundaries. The system uses template-based matching and balances privacy, latency, accuracy, and storage through explicit architectural trade-offs rather than cloud-first defaults.

Designed backend platforms that integrate ML as part of larger systems, including an end-to-end customer churn prediction pipeline and an enterprise-style RAG knowledge platform. The focus was on data flow design, retrieval accuracy vs. latency trade-offs, cost control, and traceability, treating ML as a subsystem within production architectures, not as isolated experiments.

Built and maintained Python and NPM packages used as developer tools, focusing on:

• clear APIs and documentation
• safe defaults and usability
• packaging, versioning, and distribution

These projects demonstrate attention to developer experience, backward compatibility, and maintaining software consumed by others rather than only internal code.

I choose tools based on engineering trade-offs, not preference, with a primary focus on Node.js-based backend systems:

• Node.js (NestJS) — primary backend stack for building structured, maintainable services with clear module boundaries, dependency injection, and RBAC-heavy authorization logic

• TypeScript (Next.js) — user-facing systems requiring rapid iteration, strong typing, and clear contracts with backend services

• Go — backend services where performance, concurrency, and operational simplicity are critical

• Python — data-heavy workflows, ML pipelines, and backend tooling

• Databases — PostgreSQL and MongoDB, selected based on access patterns, consistency requirements, and schema evolution needs

Systems are designed with failure modes, background processing, retries, and scalability limits in mind.

I’ve solved 500+ data structures and algorithms problems, which informs how I reason about performance, scalability, and backend design trade-offs.

Alongside engineering, I mentor as a Senior CS Instructor at Phitron, where I’ve guided 1,000+ students in core computer science and software engineering. Teaching strengthens my ability to design clear systems and communicate technical trade-offs without defining my professional identity.

• MSc in Computer Science and Engineering — University of Dhaka
  Focus areas: Quantum Computing, Artificial Intelligence & Machine Learning, Cybersecurity, and Big Data Systems

• Email: avishekdevnath@gmail.com
• LinkedIn: https://www.linkedin.com/in/avishek-devnath
• Portfolio: https://avishek-devnath.netlify.app

Currently seeking Backend Engineer / Backend-leaning Full-Stack roles where system design ownership and operational responsibility matter.