Skip to content

Portfolio · 2026

AI Developer · Full Stack Engineer

Somesh M — Portfolio

EST. 2026

Somesh M

BUILDING
Intelligent
SYSTEMS.

Computer Science student and developer building AI-powered, full-stack and intelligent software systems.

Explore workView resume

Based in India
Computer Science & Business Systems

Scroll to explore ↓

About

A little
about me.

“I build software at the intersection of artificial intelligence, full-stack engineering and product development.”

I'm a CS & Business Systems undergraduate who learns by shipping — RAG assistants, career-intelligence tools, planning platforms, and full-stack apps, from retrieval pipeline to finished interface.

I care about useful over flashy: honest scope, clean architecture, and systems that respect the people using them. Currently interning, studying, and keeping a daily problem-solving streak.

Selected work

Selected
work.

Six builds, each told as a magazine spread with its own case-study page. Hover a title; it leans closer.

SkillVision AI

Intelligent Career & Skill Intelligence System

Problem
Job seekers can't see how their skills map to real roles, and generic career advice ignores what's actually on their resume.
Solution
SkillVision parses resumes into structured profiles, scores ATS compatibility, maps skill gaps against target roles, and generates milestone roadmaps, mock interviews, and matched job leads — with RAG retrieval grounding its guidance.
Technology
React · TypeScript · Vite · Tailwind CSS · Express · Google Gemini API

My contribution — Designed and built the platform — resume parsing and ATS scoring flows, skill radar and gap logic, roadmap and interview modules, job-matching engine, and RAG-grounded career guidance.

View case studyGitHubLive Demo

OpenBlueprint

From Measurements to Intelligent Blueprints.

Problem
Early-stage home planning is guesswork: room needs, plot constraints, light, ventilation, and cost are reasoned about separately, if at all.
Solution
OpenBlueprint turns natural-language requirements into structured constraints, generates five design strategies with a zone-based BSP layout engine, and offers a CAD-like 2D editor, clay-render 3D views, INR cost estimation, and a RAG knowledge assistant.
Technology
Next.js · TypeScript · Tailwind CSS · Prisma · AI APIs · 3D Visualization

My contribution — Designed and built the platform — strategy generation, BSP-based layout logic, 2D editor tooling, 3D visualization, cost estimation, and the AI/RAG assistants.

Note — Preliminary layouts for exploration — not a substitute for a licensed architect.

View case studyGitHubLive Demo

DiagnoraX

Your Personal AI Health Companion

Problem
Health information is scattered across prescriptions, lab reports, and memory — hard to organize, easy to misunderstand.
Solution
DiagnoraX centralizes the everyday health workflow: an AI symptom checker with severity and next-step guidance, body-composition analysis, doctor recommendations, prescription OCR, interaction warnings, reminders, and report analysis — with Firebase auth and data throughout.
Technology
Flutter · Dart · React · Firebase · Firestore · Google Gemini API

My contribution — Built DiagnoraX end-to-end during my ML internship — the React web app and its Flutter conversion, Firebase auth and Firestore data, and Gemini-powered assistance flows.

Note — A software project for organizing health information — not professional medical diagnosis.

View case studyGitHub

JeduAI Connect

Smart Learning and Language Empowerment

Problem
Language barriers and manual classroom admin slow down learning in multilingual classrooms.
Solution
JeduAI Connect pairs live classes with AI assistance — attendance automation, translation, and multilingual learning paths — coordinated across student, teacher, and admin roles.
Technology
Flutter · Dart · AI APIs · Firebase

My contribution — Designed the multi-role experience and built the Flutter client with Firebase-backed classes, attendance, and translation flows.

View case studyGitHubLive DemoAndroid APK

AppleSupport AI

AI Customer Support Intelligence

Problem
Support teams drown in conversations with no systematic way to know what was asked, what worked, and what needs a human.
Solution
A retrieval pipeline over reconstructed conversations — TF-IDF baseline vs. embeddings, RAG retrieval, classification, and an explicit escalation policy, evaluated by LLMs and validated by humans.
Technology
Python · Jupyter Notebook · RAG · Embeddings · TF-IDF · LLM Evaluation

My contribution — Built the conversation-reconstruction and retrieval pipeline, classification and escalation logic, and the evaluation harness.

View case studyGitHub

[06 — Live project]

Adventure Game

“The Archive” — full-stack text adventure

A full-stack text adventure game, “The Archive,” with a Spring Boot backend, React frontend, MySQL database, and JWT authentication — three playable chapters, an achievements system, and a leaderboard. Live demo linked below.

View case study GitHub Live Demo

Experience

Where I've
been building.

  1. 01

    2026 · Internship

    Brainery Spot Technologies

    Machine Learning Intern

    ML-focused development — designed and built DiagnoraX, an AI-powered health companion, during the internship.

  2. 02

    2025 · Internship

    Nandha Info Tech

    Full Stack Development Intern

    Frontend and backend web development modules — practical experience building and shipping web applications.

05 · The foundation

B.Tech — Computer Science and Business Systems

V.S.B. Engineering College · Affiliated to Anna University · Expected graduation 2027

Skills

The tools
behind the work.

A typographic wall, not a report card. Hover any technology for context.

01

Languages

  • Python

    APIs, ML pipelines, automation

  • Java

    OOP, DSA, coursework depth

  • JavaScript

    Interactive web interfaces

  • Dart

    Flutter app development

02

AI / ML

  • Machine Learning

    Models, evaluation, iteration

  • Generative AI

    LLM-powered product features

  • RAG

    Retrieval over real knowledge

  • LLM APIs

    Gemini and friends in prod

03

Development

  • React

    Component-driven interfaces

  • Flutter

    Cross-platform mobile apps

  • Express

    Lightweight service backends

  • Firebase

    Auth, Firestore, hosting

  • Tailwind CSS

    Design systems in code

04

Tools

  • Git

    Version control, collaboration

  • Docker

    Reproducible environments

Currently deepening — RAG evaluation · Docker workflows · Secure API design

Things I've learned

Certifications.

  1. 01

    Python Programming

    Course certification

  2. 02

    Programming in Java

    NPTEL

  3. 03

    Introduction to Generative AI

    Coursera

  4. 04

    Introduction to IoT and Digital Transformation

    NASSCOM

  5. 05

    Becoming an Agentforce Champion

    Salesforce · 2025

  6. 06

    Becoming an Agentforce Champion

    Salesforce · 2026

Milestones

Built. Learned. Solved.

  • 100+

    DSA problems solved

  • 100-day

    LeetCode streak

  • 02

    Engineering internships

  • 06

    Shipped projects

  • RAG

    AI retrieval workflows

Only what's verifiable is listed — profiles linked below.

07 · Contact — © Somesh M 2026

Let's
build
something
intelligent.

Available for software engineering, AI and full-stack opportunities.

Prefer email? Email me →

Write to me

Replies land here — double-check it.