3 min readFleek
Building Fleek — A Side Project URL Shortener
What I shipped when I wanted something small, fast, and fully mine.
Side projectRedis
Read postAI-First Full Stack Engineer
AMAN
Kanojiya
Building production-ready products with Next.js, Node.js, and LLMs.
I help startups and product teams turn ideas into scalable software — combining modern full-stack engineering with AI integrations that deliver measurable business results.
About me
Engineer, builder, and problem-solver focused on shipping real products
AK4+ yrs experience
Aman Kanojiya
Software Development Engineer (Full Stack)
Over the past 4+ years, I've helped startups and businesses transform ideas into production-ready applications. I specialize in building AI-powered products using Next.js, Node.js, TypeScript, and modern cloud infrastructure. From architecting scalable systems to integrating LLM workflows, I enjoy solving complex problems that create meaningful user impact.
4+ years shipping production softwareAI / LLM integration in live productsFull-stack ownership — design to deployAWS, Azure & modern cloud stacks
Building with AI
Production LLM work — not just API calls. Real workflows shipped to users in enterprise and consumer products.
Conversational AI workflows
Designed multi-turn coaching flows with Claude and OpenAI — context-aware prompts, session memory, and production guardrails for enterprise Teams apps.
LLM integration in production
Integrated Claude and OpenAI APIs into customer-facing products with token tracking, caching, retry logic, and cost-aware prompt design.
Speech & multimodal pipelines
Built speech-to-text and voice-first UX for wellbeing coaching — Google Cloud Speech API with sentiment-aware downstream LLM flows.
AI-assisted recommendations
Shipped structured JSON recommendation engines from user behavior data — watch history, attendance, and rubric-driven session QA.
Prompt orchestration & evaluation
Domain-specific prompt engineering for edtech and workplace wellbeing, Assistants API pipelines, and audit-friendly LLM output validation.
Safety & trust layers
Rule-based URL screening plus optional OpenAI classification for user-generated links — SSRF-safe validation in production SaaS.
Work experience
4+ years delivering production impact — AI products, e-commerce, and enterprise SaaS
Software Development Engineer (Full Stack)
6DegreesIT Pvt. Ltd.
Anthropic Claude APIReactNode.jsMongoDBRedisAzureAKSMicrosoft Teams
- ›Shipped Claude-powered coaching in production for Teams — context-aware prompts, Redis session management, and token caching that reduced API cost and latency.
- ›Architected 12+ Node.js microservices on Azure Kubernetes with Service Bus and KEDA — enabling proactive wellbeing notifications at scale across time zones.
- ›Led monolith-to-microservices migration and automation workflows, improving release velocity and platform reliability for enterprise customers.
Next.js 15NestJSPrismaPostgreSQLStripeTaxCloudAWSRedis
- ›Delivered multi-vendor commerce on Next.js 15 and NestJS — enabling vendors to onboard, manage catalog, and fulfill orders through a unified platform.
- ›Implemented server-authoritative Stripe + TaxCloud checkout, eliminating client-trusted totals and reducing payment reconciliation errors.
- ›Deployed on AWS (Amplify, ECS Fargate, RDS, ElastiCache) with RBAC and CMS tooling — accelerating merchandising and order operations for admin teams.
OpenAI APIAssistants APINestJSGraphQLPostgreSQLDynamoDBAWS S3
- ›Built AI recommendation engine with structured JSON from learner behavior — personalizing practice paths and improving engagement for arts education.
- ›Automated mentor session QA via Assistants API v2 against course rubrics — reducing manual review time and standardizing feedback quality.
- ›Designed PostgreSQL + DynamoDB LLM pipelines with retry logic and deduplication — improving data reliability for downstream analytics.
Node.jsTypeScriptAI/LLMREST APIs
- ›Architected backend for concurrent AI sales simulations with low-latency scoring — supporting multiple users in production simultaneously.
- ›Redesigned KPI APIs with product and analytics teams — aligning measurable sales performance metrics with business goals.
Backend Developer (Node.js)
Otomator Technologies
Node.jsAWS EC2SQLREST APIs
- ›Developed procurement backend on Node.js — improving data access patterns and backend efficiency for enterprise workflows.
- ›Managed AWS EC2 deployment with uptime-focused operations — supporting reliable access for internal business users.
Skills & expertise
Focused expertise across full-stack product engineering, production AI, and cloud-native delivery
Core Expertise
Next.js
React
TypeScript
Node.js
NestJS
AI & LLM
OpenAI API
Anthropic Claude- Assistants API
- Prompt Engineering
- RAG
- Conversational AI
Cloud & DevOps
AWS
Azure
Docker
Kubernetes
Redis
CI/CD
Databases
PostgreSQL
Prisma
MongoDB- Redis
DynamoDB
6DegreesIT projects
Production products with measurable user and business impact — Sep 2021 to present
6DegreesIT
AI workplace wellbeing · Microsoft Teams
Joye
Problem
Organizations lacked proactive employee wellbeing insights inside Microsoft Teams — check-ins were manual and engagement dropped without timely, personalized support.
Solution
Built an AI-powered wellbeing platform embedded in Teams with Claude-driven coaching, voice-first flows, and proactive notifications across 12+ microservices on Azure.
Impact
Enabled scalable employee engagement workflows across time zones, reduced friction in wellbeing check-ins through conversational AI, and cut LLM latency costs via Redis session caching and prompt optimization.
Tech stack
Claude APIReactNode.jsMongoDBRedisAzureAKSMicrosoft Teams
ShequeMulti-vendor e-commerce platform
6DegreesIT
Multi-vendor e-commerce platform
Sheque
Problem
Multi-vendor brands needed a unified commerce stack — separate storefront, admin, and vendor tools with compliant checkout, tax, and fulfillment across sellers.
Solution
Delivered a three-app workspace on Next.js 15 and NestJS with server-authoritative Stripe checkout, TaxCloud tax compliance, RBAC, CMS, and AWS deployment.
Impact
Accelerated vendor onboarding with reusable architecture, eliminated client-trusted payment totals, and enabled faster catalog and order operations through shared admin tooling.
Tech stack
Next.js 15React 19NestJSPrismaPostgreSQLStripeTaxCloudAWSRedisBullMQ
6DegreesIT
AI practice & session quality analysis
Artfills
Problem
Edtech mentors lacked scalable ways to personalize practice and objectively QA live sessions against course-specific rubrics.
Solution
Built OpenAI-powered recommendation and Assistants API QA pipelines with GraphQL APIs, PostgreSQL + DynamoDB data flows, and admin analytics dashboards.
Impact
Automated personalized practice playlists from learner behavior and standardized mentor session reviews — improving consistency and reducing manual QA overhead for operations teams.
Tech stack
OpenAI APIAssistants APINestJSGraphQLPostgreSQLDynamoDBAWS S3
6DegreesIT
Firuta · AI sales coaching app
Sales Coach
Problem
Sales teams needed realistic practice simulations with immediate feedback — existing tools couldn't handle concurrent conversational scoring at low latency.
Solution
Architected backend for real-time AI sales simulations with concurrent user sessions, scoring pipelines, and redesigned KPI APIs aligned to product analytics.
Impact
Enabled measurable sales performance tracking and faster iteration on coaching quality through data-driven KPI definitions and real-time feedback loops.
Tech stack
Node.jsTypeScriptAI/LLMREST APIsReal-time
Personal projects
Independent products shipped to production — full ownership end to end
Personal project
Branded links platform · analytics, campaigns & integrations
Fleek
Problem
Marketers and builders sharing links needed Bitly-style tooling — short URLs, click visibility, campaigns, QR codes, and API integrations — without slow redirects or stitching together separate services.
Solution
Shipped a connections platform where redirects stay fast under load, and the dashboard covers what link teams actually use — campaigns, QR codes, bio pages, and live click charts. API keys, webhooks, OAuth apps, and custom domains sit on top so it plugs into real workflows, with optional AI screening before a bad URL goes live.
Impact
Live at app.fleek.festinme.com with go/api subdomains, sub-100ms cached redirects, GitHub Actions → GHCR → GCP deploy (Workload Identity), and Vercel-hosted web — demonstrating end-to-end product and DevOps ownership.
Tech stack
Next.jsNestJSPostgreSQLRedisPrismaOpenAI APIStripeDockerGCPVercelGitHub ActionsSwagger
Writing & notes
Short posts on things I've built — what broke, what we changed, and what I'd do again
Leave an anonymous note
No name, no email — just a message on the wall. Say hi, leave feedback, or tell me what you'd build next.
The wall
Loading notes…
Education & recognition
Education
Master of Computer Applications
Jabalpur Engineering College
Jabalpur, India
May 2021 · GPA 8.0
Object Oriented ProgrammingData Structures & AlgorithmsOperating SystemsDBMSComputer NetworksMachine Learning
Certifications
OWASP Top 10 — Web Application Security
Cybrary · May 2023
Server-side Development
Coursera · Jun 2020
Awards
Employee (Over The Top)
6DegreesIT · Sep 2022
Outstanding contributions and performance
1st Runner-Up — Smart India Hackathon (SIH) 2019
Ministry of HRD, Govt. of India · 2019
National-level hackathon recognition
Get in touch
Looking for an engineer who can take ownership, move fast, and ship production-ready AI products? Let's talk.