Stars
Commits
View on GitHub
Available for opportunities

Hardik
Parmar

Building and shipping production-grade AI systems — from serverless cloud architectures to RAG pipelines and generative applications — that solve real problems at scale.

Hardik Parmar — AI/ML Engineer

Hardik Parmar

AI / ML Engineer
Pune, Maharashtra
400+
Problems solved
1.5+
Years Experience
2024
Graduate
Introduction

About Me

I'm an AI/ML Engineer at Tata Consultancy Services, currently contributing to a defence-domain project - implementing new features and delivering AI/ML components across PI cycles in a high-stakes production environment.

I build things end-to-end. From designing ML pipelines and LLM-powered applications to deploying serverless systems on AWS that run in production — I don't just prototype, I ship. Whether it's a RAG pipeline with semantic search, a fraud detection model on imbalanced data, or a full-stack AI tracker on Lambda, DynamoDB, and Bedrock, I care about code that works under real conditions.

Driven by hard problems, clean architecture, and the craft of turning complex AI research into something a real user can actually interact with.

Production ML Systems

Designing and deploying end-to-end ML pipelines with CI/CD, model versioning, and real-world performance in mind — not just notebooks.

Generative AI & LLM Apps

Building RAG pipelines, multi-model chat interfaces, and Bedrock/Gemini-powered applications that go beyond demos into usable products.

Cloud-Native AWS Architecture

Architecting serverless systems using Lambda, DynamoDB, CDK, Cognito, and EventBridge — scalable, cost-efficient, and production-ready.

AI Agents & Intelligent Interfaces

Developing agentic systems, document Q&A tools, and AI-driven interfaces that reason over real data and deliver actionable output.

Background

Resume

Experience

Feb 2025 — current
Assistant System Engineer
Tata Consultancy Services
Those details will be updated.
Jul 2023 — Jul 2023
Data Analytics & ML Intern
Infolabz
Collaborated to optimize ML pipelines for production. Deployed image classifier with a 13% accuracy improvement using CI/CD pipelines.
Mar 2023 — May 2023
Machine Learning Intern
iNeuron.ai
Performed customer sentiment analysis across 5 platforms, enhancing targeting accuracy by 30%. Applied K-Means clustering with 27% improved accuracy. Drove 15% market share increase and 17% revenue growth.

Education

2020 — 2024
B.E. Electronics & Communication Engineering
Vishwakarma Government Engineering College, Ahmedabad
Four-year undergraduate degree in ECE with focus on ML and signal processing.
2018 — 2020
Physics, Chemistry & Mathematics
Mount Carmel High School, Gandhinagar
Completed higher secondary education with science stream.

 My Skills

Programming
PythonTypeScriptC++SQL
ML / AI
TensorFlowPyTorchKerasScikit-learnLangChainRAGLLMPandasNumPy
Generative AI
Amazon BedrockGemini APIGroqDeepSeekOpenAI API
Cloud & Infra
AWS LambdaDynamoDBS3CDKCognitoSESEventBridgeCloudFront
Databases
MongoDBMySQLDynamoDB
Dev & Deploy
FlaskFastAPIDjangoDockerKubernetesGitHub ActionsCI/CDGitLinux
Data & Analytics
TableauPower BIKafka

 Live GitHub Stats

Fetching live GitHub data…
Competitions

Hackathons

TCS & AMD AI Hackathon 2026 Top 10 finalist
Document upload
RAG retrieval
LLM validation
Scored report
Auditiq
AI-powered compliance audit validator running entirely on AMD Instinct MI300X. RAG retrieval plus an LLM validator checks documents against GDPR, SOX, HIPAA, PCI-DSS and Insurance rules, then scores and generates a full audit report with a follow-up chat assistant.
TCS's internal AI hackathon at India level — advanced through to the top 10 national finalists.
RAGFAISSQwen2.5-7BAMD ROCmFastAPI
Snapdragon® Multiverse Hackathon Runner-up · Top 8 of 50
DevMesh
On-device AI code reviewer running entirely on a Snapdragon NPU. Reviews git commits and PRs offline, streams findings to a mobile triage app, and lets developers push back on false positives before generating a local PDF report — no code ever leaves the machine.
An India-level hackathon where only 50 teams nationwide were selected for the on-site build phase in Bengaluru.
Team of 3 · Team Lead — LLM integration & backend orchestration
Snapdragon NPUQwen3-4BGenieX / QAIRTFastAPIReact Native
Git commit / PR
On-NPU review
Mobile triage
PDF report
TCS AI Friday Hackathon Finalist
Student profile
AI / logic engine
Personalized path
Personalized Learning Path Recommender
AI-powered learning-path generator with dual recommendation modes — Groq's Llama-3.3-70B for conversational guidance and a topological-sort logic engine — across 3 curriculum tracks with prerequisite mapping. Advanced to the finale as the Quadra jury's top pick; the event closed without an overall winner after a cross-location jury split.
StreamlitGroq Llama-3.3Topological Sort
Portfolio

Projects

Credentials

Certifications

Amazon Web Services
AWS Machine Learning Associate
May 2026
View credential ↗
Microsoft
GitHub Foundations
Feb 2026
View credential ↗
DeepLearning.AI
Machine Learning Specialization
March 2023
View credential ↗
Udemy
Machine Learning A-Z™: AI, Python & R
Feb 2023
View credential ↗
Google
Crash course on Python
March 2023
View credential ↗
HackerRank
Problem Solving (Intermediate) Certificate
March 2023
View credential ↗
Get in touch

Contact

Open to new opportunities, collaborations, or just a conversation about AI and ML. Feel free to reach out!