Hetvi Doshi

> AI Engineer currently

I build intelligent systems that move from prototype to production — engineering RAG pipelines, deploying LLMs, and shipping AI that actually works at scale.

~/ai-engineer — zsh
$

EXPERIENCE

Jun 2026 – Present
Surat (Onsite)

AI Engineer Intern

Active
Samniti Technologies

Built a multilingual RAG-based WhatsApp chatbot with LangChain, embeddings, and vector search for property discovery. Engineered semantic search across 2M+ listings with sub-second retrieval and optimized RAG prompting to reduce hallucinations and improve response relevance.

Python RAG PostgreSQL LangChain Embeddings Vector Database

AI/ML Intern

Completed
Wizcoder

Built an NLP-powered sales analytics assistant that converted natural language questions into database queries, enabling conversational business intelligence across multi-table datasets. Designed scalable data workflows and implemented intelligent routing for accurate real-time insights.

Python NLP LLMs FastAPI
Jan 2026 – Feb 2026
Surat (On-site)
Aug 2025
Remote

Selected Participant

Top 5% (3k/60k+)
Amazon ML Summer School 2025

Gained hands-on exposure through expert-led sessions by Amazon scientists on advanced ML topics including DNNs, LLMs, Causal Inference, and Reinforcement Learning.

Machine Learning Deep Learning LLMs
View Certificate
SIH 2025 Finalist
Smart India Hackathon
Top 3000 / 60,000+
Amazon ML Summer School
AI System Builder
AI Engineer
Python TensorFlow Flask PostgreSQL Tailwind CSS OpenAI API LangChain Docker Scikit-Learn Git & CI/CD
ENGINEERING PROOF

My Projects

Not just demos. Production-grade AI systems engineered to solve complex, real-world problems with measurable impact.

SIH 2025 Finalist

AI-Based DPR Evaluation System

An AI-powered system that automates the evaluation of Detailed Project Reports (DPRs) by analyzing compliance, risks, and guideline adherence. Built using NLP, hybrid PDF extraction, and RAG to process multilingual and scanned documents, reducing manual review efforts from hours to minutes.

Python NLP RAG FastAPI
Architecture Flow
PDF Input Hybrid RAG Rule Engine
Multi-Modal AI

NURA Wellness App

A multi-modal AI wellness platform that analyzes behavioral, wearable, Wi-Fi, and conversational data to assess mental health risks. Leverages LangGraph workflows and LLM-powered reasoning to generate personalized wellness insights and automated reports.

Python FastAPI Langchain Langgraph
Architecture Flow
Sensors LangGraph Risk Score
NL-to-SQL Agent

Slack AI Data Bot

An AI-powered business intelligence assistant that allows users to query databases using natural language directly from Slack. Converts user questions into SQL queries, retrieves relevant insights, and delivers real-time analytics through a conversational interface.

Python FastAPI Langchain PostgreSQL
Architecture Flow
Slack API NL-to-SQL Postgres
Amazon ML Challenge 2025

Multimodal Product Price Prediction

Developed a multimodal ML pipeline for product price prediction using image, text, and structured data. Leveraged ResNet50V2, TF-IDF, and LightGBM to extract meaningful features from noisy catalog data and achieved a top 25% leaderboard ranking in Amazon ML Challenge 2025.

Python LightGBM ResNet50V2 TF-IDF Machine Learning
Pipeline Architecture
Images Text Structured Feature Extraction LightGBM Price Pred
GitHub Analyzer

Code Portfolio Evaluation Agent

A GitHub-based portfolio evaluation platform that analyzes repository quality and engineering signals to generate explainable scores, insights, and actionable improvement recommendations.

Python React FastAPI Next.js TypeScript
Evaluation Pipeline
GitHub Analyzer Score
ENGINEERING STACK

Tools & Technologies

Core Languages

Primary languages for data processing, backend logic, and database querying.

Python SQL

AI / Machine Learning

Techniques and paradigms for building intelligent, data-driven systems and pipelines.

Machine Learning NLP LLMs RAG Feature Engineering Model Evaluation

Frameworks & DBs

Robust, typed, and scalable server architectures paired with reliable data storage.

FastAPI PostgreSQL SQLite

Data & ML Libraries

Essential Python ecosystem packages for data manipulation, visualization, and statistical modeling.

Pandas NumPy Scikit-learn NLTK Matplotlib Seaborn

Developer Tools

Git • GitHub • Docker • Jupyter Notebook • Google Colab • VS Code

Achievements & Credentials

Finalist 2025

Smart India Hackathon

SIH 2025 Finalist • National Level

Selected 2025

Amazon ML Summer School

Selected: 0 / 60,000+

Semi-Finalist 2026

The Economic Times GenAI Hackathon

Semi Finalist • High-scale AI problem-solving

Ranked 2025

Amazon ML Challenge

Rank 0 / 6,000+ Teams

Finalist 2026

Odoo x KSV Hackathon '26

Finalist: Top 0 / 2,000+ Teams

Verified Credentials

OCI Data Science Professional

Oracle University

View Credential

IBM SkillsBuild - Data Analytics

CSRBOX × IBM

View Credential

Data Analysis and AI/ML

SAP × Edunet

ENGINEERING METHODOLOGY / 001

From concept to production.

A disciplined, six-stage approach to designing, building and operating intelligent systems. Every decision grounded in performance, reliability and real-world impact.

ENGINEERING PIPELINE
PROCESS STATUS SYSTEM READY
PHASE 01 / 06 RESEARCH & DISCOVERY

Discover the problem.

Understand the business objective before choosing the technology. Audit data quality, identify constraints and define measurable success criteria.

Data Audit Requirements Feasibility
KEY DELIVERABLE

Validated requirements and success metrics

WORKFLOW PROGRESS 01 / 06
01 / PRINCIPLE

Measure what matters.

Define clear performance targets before engineering begins.

02 / PRINCIPLE

Design for scale.

Build modular systems that remain maintainable as requirements evolve.

03 / PRINCIPLE

Engineer for reliability.

Validate, monitor and improve systems throughout their lifecycle.

System Uplink

Have a complex
engineering problem?
Let's build the solution.