AI Developer building production-grade AI systems — LLM apps, RAG pipelines, autonomous agents, and full-stack AI products that actually ship. Python · FastAPI · React. MNIT Jaipur, B.Tech 2025.
I’m Kartikey Sharma — an AI developer who ships production-grade AI systems (LLM apps, RAG, agents, full-stack), not throwaway demos. I help teams automate workflows, build intelligent products, and deploy AI that stays reliable in the real world. MNIT Jaipur background; currently open to freelance and serious collaborations.
Short intro — who I am and how I work.
Enterprise-grade AI platform that autonomously routes, triages, and resolves customer support tickets. Built with LangGraph state machines coordinating specialized triage, knowledge retrieval, and resolution agents with full conversation memory and context awareness.
AI-powered workflow automation platform that runs 100% locally using open-source models. Features autonomous agents for research, analysis, content creation, visual workflow builder, vector memory, and event-driven automation — all without API costs or data privacy concerns.
Full-stack personal AI assistant with real-time chat via WebSocket, intelligent reminders, alarms with push notifications, and a beautiful Progressive Web App interface. Features Firebase Cloud Messaging, activity dashboard, and quick actions for daily productivity.
Intelligent financial analysis platform that transforms raw financial data into actionable insights. Features automated spending pattern analysis, anomaly detection, budget recommendations, and beautiful visual dashboards for comprehensive financial intelligence.
Multilingual AI travel companion with GPS-powered location discovery, personalized itinerary recommendations, and persistent conversation memory. Features Groq + Gemini fallback for reliable responses, Google Places integration, and seamless travel planning experience.
Revolutionary natural language to production-ready code system. Multi-agent architecture with Planner, Architect, and Coder agents working together to simulate real software development workflows — from concept to deployment-ready applications.
Advanced AI research assistant that autonomously conducts comprehensive research, analyzes multiple sources, synthesizes findings, and generates detailed research reports. Features intelligent source validation, citation management, and structured output formatting.
Comprehensive neural network framework and experimentation platform. Build, train, and deploy deep learning models with advanced architectures, hyperparameter optimization, and production-ready model serving capabilities.
Intelligent data management and analytics platform that drives actionable insights from complex datasets. Features automated data pipelines, real-time processing, advanced analytics, and visualization tools for data-driven decision making.
AI-powered review management and analysis system that intelligently processes, categorizes, and extracts insights from customer reviews. Features sentiment analysis, trend identification, and automated response generation for enhanced customer engagement.
A four-service EdTech platform behind a single SSO gateway — a Portal reverse-proxies role-based Student, Teacher, and Admin apps. Spans a grade-adaptive study companion, an AI teacher toolkit (~50 FastAPI endpoints, an MCP server, and adaptive-learning ML with IRT + Bayesian Knowledge Tracing), and a governance dashboard with dual-DB analytics and an early-warning student-risk system.
An AI "control room" that replaces a school's paper registers and phone calls with one dashboard. Features a from-scratch timetable constraint solver (120 lessons in ~30 ms), an instant substitute finder, a discipline pattern engine, and a smart circular system with real WhatsApp/SMS delivery and read-receipt tracking.
A voice-first AI tutor with a lip-synced 3D talking avatar that teaches the CBSE/NCERT K-12 syllabus Socratically — in English, Hindi, and Hinglish. Real-time STT→LLM→TTS over WebSocket, a live whiteboard, doubt-from-photo vision, diagnostic prerequisite gap-mapping, and mastery tracking with spaced repetition.
An AI autograder for scanned, handwritten CBSE Class-10 Science copies — proven at 96.4% agreement within ±1 mark on a 50-copy golden set. A teacher-approval rubric gate plus a deterministic override layer (sympy re-checks numerics, chempy verifies balanced equations) means the LLM is never trusted for math.
A full-stack bulk answer-sheet grader (up to 50 sheets, any grade or subject) built on a LangGraph state-machine pipeline. Teacher instructions become hard LLM constraints, a verifier agent flags over-generous marks, and layered OCR degrades gracefully across Gemini, Groq, and Google Cloud Vision.
An AI medical copilot whose hero feature analyzes 30+ lab & imaging report types via GPT-4.1 vision. A cost-efficient hybrid runs heavy vision in the browser (Puter.js) while lightweight NLP stays server-side — plus doctor/medicine recommendations, multilingual health chat, and real-time AQI/pollen alerts.
An AI tutor-matching agent that turns a 3–4 hour spreadsheet-filtering workflow into a ~2-minute human-in-the-loop process. It scans a 3,000+ teacher pool and ranks candidates with a transparent weighted 0–100 scoring model using Haversine distance for home-tuition proximity — with a readable "Why this match?" breakdown.
An AI financial analyst that ingests filings, extracts the three financial statements into validated data, and writes cited investment memos. Core principle — numbers come from code, narrative from the model: deterministic ratio engines plus forensic screens (Altman Z, Beneish M, Piotroski F) eliminate arithmetic hallucination.
A motion-engineering showcase for evidence-first investigative journalism. Bespoke interaction craft — an AI text-decrypt headline, a mouse-reactive neural-constellation canvas, a declassify spotlight, and a spring-trailing dual cursor that morphs into labels — on a bleeding-edge front-end stack.
A stylish two-wheeler parts storefront (petrol + EV): pick a vehicle and animated side-profile art reveals fitment hotspots synced to a categorized parts panel with live pricing and compatibility scores — plus a build/cart drawer handling quantities, services, GST, and checkout.
Custom autonomous agents using LangGraph — ReAct, self-reflection, task decomposition. From single tools to complex multi-agent orchestration with memory.
Replace manual processes with intelligent AI pipelines. Customer support, document processing, research workflows — running reliably without human intervention.
Intelligent Q&A over your documents using vector embeddings and ChromaDB. Understands context, not just keywords.
React frontend + FastAPI backend + Docker. Real-time streaming, beautiful UIs, production-ready. Full product, not just an API wrapper.
Structured architectures, JSON enforcement, hallucination reduction, guardrails. Making LLMs reliable and consistent in production.
Supervised/unsupervised models, feature engineering, intelligent dashboards. Raw data into actionable business intelligence.
Deep dive into your workflow. What's manual? What costs time? I map the exact process before writing a single line of code.
Design the LangGraph state machine — agents, tools, handoffs, data flow. The right architecture determines everything.
Real LLM integrations from day one. Weekly demos, constant feedback, no surprises. You see progress every step.
Docker, error handling, health checks, documentation. Not a demo — a system your team can rely on every day.
"I've worked with Kartikey on a few AI projects and he's genuinely good. Clean, well-structured code and a real understanding of agent architectures — not just API wrappers. Ships fast and is easy to build with."
"Good experience overall. The research agent saves us time on desk research. We still tweak it sometimes for niche topics, but the base system is strong."
"Understood our requirements quickly and shipped on the agreed timeline. Some edge cases came up later, but he helped us handle them without drama."
"The finance tool is now part of our workflow. UI is clean, and streaming responses feel nice. There were a couple of minor bugs at launch, but they were fixed fast."
"Knows his way around production AI. Not just fancy demos. The RAG system is reliable enough that our team actually uses it daily."
"Worked alongside Kartikey on agent and automation systems. He really knows his way around AI agents and harness engineering, and the work held up in production. Sharp, reliable, and someone I'd happily collaborate with again."
"I've worked with Kartikey on a few AI projects and he's genuinely good. Clean, well-structured code and a real understanding of agent architectures — not just API wrappers. Ships fast and is easy to build with."
"Good experience overall. The research agent saves us time on desk research. We still tweak it sometimes for niche topics, but the base system is strong."
"Understood our requirements quickly and shipped on the agreed timeline. Some edge cases came up later, but he helped us handle them without drama."
"The finance tool is now part of our workflow. UI is clean, and streaming responses feel nice. There were a couple of minor bugs at launch, but they were fixed fast."
"Knows his way around production AI. Not just fancy demos. The RAG system is reliable enough that our team actually uses it daily."
"Worked alongside Kartikey on agent and automation systems. He really knows his way around AI agents and harness engineering, and the work held up in production. Sharp, reliable, and someone I'd happily collaborate with again."