Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Backend engineer and startup operator building production, AI-integrated systems end-to-end. Currently Backend Engineer & CMO at Acadex, architecting scalable Node.js/MongoDB backends and routing LLM inference (NVIDIA NIM, Groq) into deterministic, policy-enforcing workflows for 500+ concurrent users. I’ve shipped an e-commerce platform and an AI SaaS product end-to-end—owning auth, catalog, order flows, inference pipelines, and output-validation layers to make non-deterministic models reliable at the application layer. Comfortable with founder-level autonomy across backend architecture and go-to-market.…Backend engineer and startup operator building production, AI-integrated systems end-to-end. Currently Backend Engineer & CMO at Acadex, architecting scalable Node.js/MongoDB backends and routing LLM inference (NVIDIA NIM, Groq) into deterministic, policy-enforcing workflows for 500+ concurrent users. I’ve shipped an e-commerce platform and an AI SaaS product end-to-end—owning auth, catalog, order flows, inference pipelines, and output-validation layers to make non-deterministic models reliable at the application layer. Comfortable with founder-level autonomy across backend architecture and go-to-market.

Shaan Goswami

Back-End Developer, Full Stack Developer, AI Developer, +7





Backend engineer and startup operator building production, AI-integrated systems end-to-end. Currently Backend Engineer & CMO at Acadex, architecting scalable Node.js/MongoDB backends and routing LLM inference (NVIDIA NIM, Groq) into deterministic, policy-enforcing workflows for 500+ concurrent users. I’ve shipped an e-commerce platform and an AI SaaS product end-to-end—owning auth, catalog, order flows, inference pipelines, and output-validation layers to make non-deterministic models reliable at the application layer. Comfortable with founder-level autonomy across backend architecture and go-to-market.…Backend engineer and startup operator building production, AI-integrated systems end-to-end. Currently Backend Engineer & CMO at Acadex, architecting scalable Node.js/MongoDB backends and routing LLM inference (NVIDIA NIM, Groq) into deterministic, policy-enforcing workflows for 500+ concurrent users. I’ve shipped an e-commerce platform and an AI SaaS product end-to-end—owning auth, catalog, order flows, inference pipelines, and output-validation layers to make non-deterministic models reliable at the application layer. Comfortable with founder-level autonomy across backend architecture and go-to-market.

Available to hire

Backend engineer and startup operator building production, AI-integrated systems end-to-end. Currently Backend Engineer & CMO at Acadex, architecting scalable Node.js/MongoDB backends and routing LLM inference (NVIDIA NIM, Groq) into deterministic, policy-enforcing workflows for 500+ concurrent users.

I’ve shipped an e-commerce platform and an AI SaaS product end-to-end—owning auth, catalog, order flows, inference pipelines, and output-validation layers to make non-deterministic models reliable at the application layer. Comfortable with founder-level autonomy across backend architecture and go-to-market.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Beginner

Beginner

Language

Work Experience

Developer (AI SaaS platform) at ColdMailAI

January 1, 2025 - December 31, 2025

Built a production-grade AI SaaS platform routing LLM inference through NVIDIA NIM (nemotron-ultra-550b, llama-3.3-70b-instruct) and Groq API. Designed backend policy logic to validate, filter, and enforce structured outputs from generative models and architected the full end-to-end system including auth, prompt pipeline, API gateway design, and rate limiting.

Founder-style full-stack developer (e-commerce) at JuteIt

January 1, 2025 - December 31, 2025

Built a full-stack e-commerce platform end-to-end including authentication, product catalog, and order workflows with payment flows. Implemented RESTful APIs for product management, order processing, and checkout with Firebase Authentication and role-based route protection. Deployed to production with consistent sub-500ms load under concurrent traffic.

Backend Engineer & CMO at Acadex by Ereefian

January 1, 2024 - Present

Architected and integrated Groq-based LLM inference pipelines into backend workflows, adding response validation and output-enforcement logic to keep non-deterministic model behavior reliable at the application layer. Shipped 6+ scalable backend modules for an AI-powered academic management platform, improving average API response time by ~35% via query optimization and schema redesign. Owned system design for role-based access control, event management, and notification infrastructure, reducing feature release cycles by ~40% through modular extensible architecture. Built REST APIs and MongoDB schemas plus Firebase-integrated workflows handling institutional data for 500+ concurrent users under real production load. Led go-to-market and platform expansion strategy as CMO with engineering ownership using founder-level autonomy.