Portrait of Samarth M NSamarth M NSenior AI Full-Stack Engineer

Senior AI Full-Stack Engineer

I build reliable AI products, end to end.

Applied LLMs, RAG, and AI Products

I’m a senior full-stack and product engineer with 8+ years of experience shipping software across SaaS, healthcare, education, Web3, and consumer products. My current focus is applied LLM systems built with Python, FastAPI, React, Next.js, TypeScript, local models, and commercial model APIs.

View RAGnosisExplore ProjectsView Résumé

Open to senior AI full-stack, LLM application, founding-engineer, and Forward Deployed Engineer opportunities.

8+ yearsShipping product software
0.406 → 0.792RAG retrieval MRR across tracked experiments
Founder-level ownershipFrom product discovery to deployment
Toptal network experienceRemote client and stakeholder delivery

What I bring

Senior product engineering experience, applied to modern AI systems.

My background is in building complete products: translating ambiguous requirements into architecture, shipping frontend and backend systems, integrating external services, handling authentication and permissions, and carrying features through deployment.

I now apply that experience to LLM applications. That includes retrieval pipelines, model routing, structured API boundaries, evaluation workflows, local models, multimodal interfaces, and the React or Next.js experiences through which users interact with the system.

My advantage is not limited to prompting a model. I can work across the full product surface surrounding it.

LLM Application Engineering

Build retrieval workflows, model integrations, structured outputs, model-routing layers, evaluation pipelines, local-model tools, and multimodal AI experiences.
RAGembeddingsvector searchrerankingcontext engineeringmodel evaluationmodel routinglocal LLMsmultimodal AI

AI Full-Stack Product Development

Build the API, data layer, authentication, frontend architecture, product interface, analytics, monitoring, and deployment around an AI workflow.
PythonFastAPITypeScriptReactNext.jsNode.jsPostgreSQLSupabaseAPI integration

Founding and Forward-Deployed Delivery

Work directly with founders, stakeholders, or customers to define the problem, design the workflow, integrate external systems, resolve delivery constraints, and carry the result into a usable product.
Solution designProduct discoveryStakeholder collaborationRapid prototypingIntegrationDeploymentTechnical ownership

From an ambiguous problem to a deployed product.

  1. 01

    Understand the workflow

    Clarify the user, business objective, data, constraints, failure conditions, and how success will be evaluated.

  2. 02

    Build across the stack

    Connect model or retrieval workflows to typed APIs, data systems, authentication, observability, and a usable product interface.

  3. 03

    Evaluate and harden

    Test outputs, document limitations, handle provider and tool failures, improve reliability, and measure the behavior that matters.

Featured projects

Evidence before adjectives.

View all projects
Multimodal Agentic Cooking Assistant

Cook Mantra

An application that turns an ingredient photo into complete recipes through multimodal and specialized LangGraph agents for extraction, dish suggestions, nutrition estimates, preview images, and recipe generation.

Next.jsReactTypeScriptPythonFastAPILangGraphOllama
Source code
Applied RAG Evaluation

RAGnosis

An eight-run benchmark comparing retrieval and reranking strategies over a synthetic clinical database. The project evaluates how query rewriting, small-to-big retrieval, aggregate rollup documents, and rerankers affect retrieval and answer quality.

0.406 → 0.792 retrieval MRRBest documented answer-quality score: 4.622/5
PythonChromaOllamaOpenAI-compatible embeddingsHugging FaceBGE RerankerJina RerankerEvaluation pipelines
Interactive reportSource codeCase study
Full-Stack LLM Routing Application

Pretendo

A deployed Next.js and FastAPI LLM application with model routing, validated API boundaries, backend-owned prompts, provider timeouts, browser-side persistence, analytics, automated checks, and Vercel deployment.

Next.jsTypeScriptFastAPIPydanticOpenRouterSQLiteVercel
Live applicationSource codeCase study

Work in progress

Currently Building

In Progress

Fine-Tuning Evaluation Project

An in-progress project comparing a base model with a LoRA or QLoRA adapter on a defined task, including dataset documentation, training configuration, evaluation, error analysis, and deployment trade-offs.
In Progress

Multi-Agent Orchestration Project

An in-progress system focused on typed tools, stateful orchestration, retry and repair paths, human approval, tracing, task evaluation, and measurable reliability.

Experience building products across the stack.

Before focusing on applied LLM systems, I spent more than eight years delivering SaaS, healthcare, education, Web3, mobile, and consumer products. That experience shapes how I approach AI applications: as complete systems that require product judgment, secure data handling, maintainable architecture, and reliable delivery.

View complete résumé

DocKosha

Founder and Founding Engineer · Full-time until Jun 2026 · Passive since Jun 2026

Gondola AI

Full-Stack Engineer

Micro

Full-Stack Development Engineer

Toptal Network

Senior Product Engineer, selected client engagements

HealthPlix Technologies

Senior Software Engineer

Cloudnine Group of Hospitals

Full-Stack Engineer

Contact

Building an AI product that needs more than a model call?

I’m interested in senior AI full-stack, LLM application, Forward Deployed Engineer, AI product, and founding-engineer opportunities where I can own the path from an ambiguous problem to a deployed product.

Email SamarthView RésuméConnect on LinkedIn