Resume

Jai Kumar Meena

AI/ML Engineer — Agentic Systems & Tool-Use (A2A, MCP), LLMs/SLMs, Edge/Cloud

New Delhi, India
JK422331@gmail.com
+91 8800682416

Build and ship agentic AI applications with multi-tool orchestration (LangChain, MCP, A2A patterns), retrieval, function calling, and safety guards. Production experience across mobile, web, and backend (Python, TypeScript, FastAPI, Flutter) with vector search, caching, and observability. Hands-on with Gemini, OpenAI, SentenceTransformers; deploys TfLite/CNN on-device; integrates Vision, Maps, and speech for multimodal UX. Seeking to accelerate agentic AI and developer platform efforts across tool-using agents, inference optimization, and end-user apps.

Experience

Data Architecture Design Specialist (Contract)

US-based Enterprise AI Startup

Apr 2025 - Present
  • Designed agentic RAG system with tool-use and function-calling chains for compliance gap analysis.
  • Orchestrated via LangChain, MCP-style tool interface definitions, and structured outputs; reduced analyst review time by ~40%.
  • Implemented vector search (FAISS/Chroma) with hybrid retrieval and reranking (SentenceTransformers); added eval harness (faithfulness, groundedness, latency SLAs).
  • Built FastAPI microservices with streaming tokens, retry/backoff, circuit breakers, Redis caching; cut P95 latency from 1.6s→700ms.
  • Frontend in React/Tailwind/Shadcn with audit trails and prompt versioning; RBAC over Firebase + Postgres.
  • Integrated Gemini 2.0 Flash for rapid tool-use; added small local SLM (e.g., Phi-3/LLama-3.2 3B) for on-device redaction and cost control; automatic model routing by task/latency budget.

AI & Full-Stack Developer (Contract)

Luxury Rydes

May 2025 – July 2025
  • Delivered tri-platform apps (iOS/Android/Web) in 12 weeks using Flutter + Firebase; provisioned Cloud Functions for realtime pricing, driver ETA, and surge logic.
  • Built support agent with retrieval over policy/FAQ and tools for booking lookup, refund quotes, and route estimation; live handoff to human via ticket API; reduced median resolution time by ~35%.
  • Developed Maps APIs + geo-indexed queries; added anomaly detection for fraudulent trips using heuristic+embedding features.

AI Developer (Part-Time)

IrisIdea TechSolutions

Apr 2024 – Present
  • Promoted from intern to lead 'Leaf Aid' app; implemented custom plant-disease CNN converted to TfLite with quantization for edge inference; added offline first design and secure auth.
  • “Hibiscus” chatbot: Gemini tool-use for diagnosis, care schedule, and treatment lookups; integrated safety filters and disclaimers; >20k sessions, crash-free rate >99.6%.
  • Set up telemetry (Firebase Crashlytics, custom events), A/B prompts, and rollout via staged tracks on Play Store.

AI Model Trainer (Freelance)

Outlier

Sep 2024 – Nov 2024
  • Authored high-difficulty code and math prompts; applied RLHF-style preference ranking; specialized prompt/program-of-thought for tool-former-style usage.
  • Evaluated chain-of-thought suppression vs structured rationales; improved exact-match on held-out tasks by ~3–5pp.

Technical Skills

ai Ml

Agentic AI: LangChain, LangGraph, LlamaIndex, AutoGen, CrewAIModels: Gemini 1.5/2.0, Claude 3.5 Sonnet, GPT-4o, Llama-3, Phi-3Frameworks: TensorFlow, PyTorch, Scikit-learn, OpenCVRetrieval: RAG, Vector DBs (Chroma, FAISS, Pinecone), Hybrid SearchTools: HuggingFace, Ollama, MCP (Model Context Protocol)

frontend

React, Next.js 14, TypeScriptTailwind CSS, Shadcn UI, Framer MotionThree.js, React Three FiberState Management: Recoil, Zustand, React QueryStreamlit (Python UI)

backend

Python (FastAPI, Flask)Node.js, ExpressDatabases: PostgreSQL, Redis, MongoDB, FirestoreCloud: Google Cloud Platform (GCP), Firebase FunctionsAuth: NextAuth, Firebase Auth, JWT

mobile

Flutter (Dart) - Cross-platformNative Android (Kotlin/Java knowledge)Mobile AI: TensorFlow Lite, On-device inferenceGoogle Maps IntegrationState Management: Provider, Riverpod, Bloc

languages

PythonTypeScript/JavaScriptDartC++SQL

tools

GitHub, VS Code, PostmanOpenCV, MediaPipe, pyttsx3Docker, Kubernetes, CI/CD

Selected Projects

Smart AI Agent IDE

Full-Stack AI-Powered IDE

Python, Streamlit, Google Gemini AI, NetworkX

Built hierarchical/graph planner with memory and tool registry (HTTP, SQL, file, search); supports MCP-compatible JSON tool schemas and function-call I/O. Added task decomposition heuristics, reflection, and retries; improved task success on internal benchmarks by ~25–40% vs vanilla single-step prompting. Implemented run logs, traces, and cost/latency budgets per subtask; pluggable retrievers and safety rules.

Just-Office

Full-Stack Web App

React.js, Node.js, Express.js, MongoDB

Built a web application to manage office spaces and bookings, reducing manual allocation time by 40%. Designed a seamless UI, improving user interaction satisfaction scores, with a robust backend handling 1,000+ daily requests.

Smart Teaching Assistance Application

Web App

Python, Streamlit, Google Gemini AI

Engineered an AI teaching assistant, automating lesson plan generation and saving educators up to 30% of preparation time. Included an interactive Q&A system, supporting 95% accurate responses and enhancing educator productivity by 20%.

YT Note Card Generator

Web App

Python, Streamlit, GenAI, YT Transcript API, G-Cloud

Created a website using Google Gemini Pro for note card generation from YouTube videos. Achieved 100% accuracy in generating 10-12 insightful note cards per video, enhancing user learning efficiency by 25%.

Real Estate Assistant

Web App

Python, Streamlit, Google Generative AI, pandas

Produced an AI chatbot for real estate queries, improving user engagement by 30%. Integrated PDF and CSV processing for property data, reducing query response time by 40%.

AIEndoEnforcer

Application

Python, OpenCV, MediaPipe, Streamlit, Google Gemini AI, pyttsx3

Developed an AI-powered workout assistant for real-time exercise tracking using computer vision and pose estimation. Real-time rep counting and posture scoring; voice feedback; designed for low-latency edge inference.

Education

BTech (Computer Science Engineering)

Delhi Technological University, New Delhi

2025CGPA: 7.0

Grade XII

Kendriya Vidyalaya, New Delhi

202176.6%

Grade X

Kendriya Vidyalaya, New Delhi

201973.6%

Achievements

Amazon ML Hackathon

Ranked top 50 out of 78,000 Participants

SIH 2024 Finalist

Selected as a finalist in all-India competition from 86,000 participants

Vihaan007 Hackathon

Ranked top 25% out of 390 teams, advanced to final round

Science Olympiad

1 Gold and 2 Bronze medals