Chinmay Vivek, Applied AI Research Engineer in Bangalore
Applied AI Research Engineer

Chinmay
Vivek

Shipping Scalable AI Products for Startups

Generative AI Engineer

View portfolioBook a call
  • LLM Systems
  • RAG Pipelines
  • AI Agents
  • MLOps
  • FastAPI
  • Kubernetes
IIT JodhpurProduct AI for startupsBangalore, India
About

Product builder

Applied AI engineer shipping scalable products for startups — LLM apps, RAG systems, and cloud-native backends from Bangalore.

How I operate

I am Chinmay Vivek, an Applied AI Research Engineer focused on startup products. My work spans generative AI, LLM applications, retrieval augmented generation (RAG), multi-agent systems, and AI infrastructure, including model tuning, FastAPI services, and Kubernetes-based MLOps on AWS and GCP.

What I ship

  • Ship MVPs and full-stack products for early-stage startups
  • Design cloud-native backends in Go, Rust, and Python
  • Integrate LLMs, generative AI, and computer vision into live products
  • Move ML prototypes from proof of concept to production on Kubernetes

Core strengths

AI and machine learning82%
System architecture85%
Programming88%
PythonGolangPyTorchFastAPIKubernetesAWS
Trusted by

Trusted by

Product and platform teams across security, AI, fintech, and consumer — contract and full-time.

Portfolio

Selected work

AI systems and full-stack products in production: computer vision, LLM stacks, fintech ML, and cloud-native platforms.

SumeruFeatured

ERP

Sumeru

An open-source, modular ERP framework in Go. Every business function — contacts, products, CRM, sales, invoicing — is an addon: a folder with a manifest, Go models, and XML. The kernel discovers it, resolves dependencies, creates tables, and wires menus. PostgreSQL underneath, server-rendered HTML on top — no front-end build step.

  • Deployment is a copy: one Go binary beside Postgres — no workers, brokers, or upgrade theatre
  • Customise without forking core — three tiers (framework, standard addons, your workspace)
  • Plain Postgres tables you can query; events, ACLs, JSON-RPC, and multi-company built into the kernel
GoPostgreSQLXMLOpen Source
Visit site
SentinelVisionComputer Vision

SentinelVision

Real-time person and object tracking for production vision pipelines, built for speed and reliability in Rust.

RustOpenCVONNXSORT
View project
AI-Powered Trading Platform
FinTech AI

AI-Powered Trading Platform

Algorithmic trading stack with ML market signals, FastAPI services, and Redis for fintech automation.

PythonTensorFlowFastAPIRedis
ML Risk Assessment System
FinTech

ML Risk Assessment System

Credit scoring and fraud detection service for banking workflows, from model inference to containerized delivery.

Rustscikit-learnPostgreSQLDocker
Enterprise Mobile Suite
Full Stack

Enterprise Mobile Suite

Cross-platform field app for data capture and analytics with Flutter, Firebase, and GraphQL.

FlutterDartFirebaseGraphQL
Advanced LLM SystemGenerative AI

Advanced LLM System

Custom LLM training and serving stack for enterprise NLP and generative AI workloads on AWS.

PythonPyTorchCUDAAWS
Cloud-Native ML Platform
MLOps

Cloud-Native ML Platform

Kubernetes ML platform for automated training, deployment, and monitoring with Go, MLflow, and Terraform.

GoKubernetesMLflowTerraform
Crypto Trading Bot
FinTech

Crypto Trading Bot

Automated crypto trading bot in Go with real-time market analysis, PostgreSQL, and Redis.

GoDockerPostgreSQLRedis
Real-time Analytics DashboardFull Stack

Real-time Analytics Dashboard

Streaming analytics dashboard in Rust and Next.js for live operational visibility.

RustNext.jsPostgreSQLRedis
Live ClassFull Stack

Live Class

Open source LMS for live online classes: courses, attendance, and real-time sessions with Go and HTMX.

GoDockerPostgreSQLHTMX
View project
Swift School
Full Stack

Swift School

School administration system for admissions, academics, fees, and reporting on Go and PostgreSQL.

GoDockerPostgreSQL
View project
Stack

Technical stack

The systems stack behind the products: generative AI, LLM engineering, RAG, agents, MLOps, and distributed ML on AWS, GCP, and Kubernetes.

Product & systems

Backend & Systems

GoRustPythonMicroservices

Enterprise Solutions

ERP SystemsFinTech SolutionsPayment SystemsBanking APIsOdoo

Frontend & Mobile

Next.jsFlutterReactHTMX

AI & ML

PyTorchTensorFlowLLMsRAGLangChainLlamaIndexFastAPIMLOps

Cloud & delivery

Cloud Platforms

AWSGCPLambdaGKECloud Run

Infrastructure & Orchestration

KubernetesDockerTerraformIaC

Architecture Patterns

MicroservicesEvent-Driven Architecture

AI systems lab

Deep learning, LLMs, and knowledge systems in production

LLM & Model Architecture

PEFTLoRAQLoRATransformers

RAG & Knowledge Processing

Vector StoresSemantic SearchKnowledge Graphs

Agent Architecture

Multi-Agent SystemsLangChainTool Use

Model Operations

Model ServingA/B TestingMonitoring

Generative AI

RAGVector DBAI Agents

MLOps & Deployment

CI/CDKubernetesDocker

Cloud Native

KubernetesDockerGitOps

Distributed Systems

MicroservicesgRPCGraphQL

Core stack

PythonGolangRustPyTorchTensorFlowDockerKubernetesAWS

Need LLM or MLOps leadership on your product? Book a short intro.

Book a call
Lab

Applied lab

Where product meets research: generative AI, LLM systems, RAG, agents, and production ML for fintech and enterprise.

AI and machine learning

15+ projects

Focus areas

  • Deep learning model design
  • Natural language processing (NLP)
  • Computer vision for production apps
Lab impact15+ shipped projects

Generative AI and LLMs

10+ projects

Focus areas

  • Large language model (LLM) fine-tuning
  • Text-to-image and multimodal pipelines
  • Prompt engineering and evaluation
Product impact12 enterprise LLM deployments

Enterprise and cloud AI

20+ deployments

Focus areas

  • Distributed systems on Kubernetes99.9% uptime
  • Multi-cloud AI infrastructureAWS and GCP
  • Microservices for ML APIs1M+ API calls
Business impact50+ client engagements
Contact

Let's build

Open to product partnerships with startups and teams shipping AI systems. Based in Bangalore.

Book a call

Prefer a direct conversation? Hold 30 minutes — pick a slot that works.

Schedule a meeting