MeoUo

Prompt

Based on the given resume wha questions am i most likely to be asked in a inteview Education M.G.M.’s Jawaharlal Nehru Engineering College (JNEC), MGM University 2024 – 2027 B.Tech in Artificial Intelligence & Data Science | Minor: Cybersecurity CGPA: 8.49 / 10 Technical Skills Languages: Python, Java, JavaScript, C/C++, SQL CS Fundamentals: Data Structures & Algorithms, OOP, DBMS, Operating Systems, Computer Networks, System Design Backend & DevOps: FastAPI, Spring Boot, REST APIs, Webhooks, PostgreSQL, Redis, Docker, Git, GitHub Actions (CI/CD), Linux AI & GenAI: LangGraph, LangChain, RAG, pgvector, OpenAI Decisions API, Jev (TypeSafe), PyTorch, Hugging Face, Scikit-learn, XGBoost, LightGBM, Pandas, NumPy Cloud: Oracle Cloud Infrastructure (OCI), Microsoft Azure, AWS, Google Cloud (Vertex AI), Railway, Vercel Experience GIS Intern 2023 Flyview GIS Technology Pvt. Ltd – GeoTech Institute Chh. Sambhajinagar, India – Structured 500+ satellite-imagery records into a relational dataset across 3 client projects, following each client’s data standards. – Wrote rule-based validation checks that caught entry and coordinate errors, improving data accuracy by 30%. Projects Recova – Agentic Payment Recovery | Python, LangGraph, Jev, FastAPI, Razorpay MCP 2026 – Built an AI agent on a LangGraph state machine that ingests signature-verified Razorpay webhooks with idempotency, classifies each failed payment, and runs a bounded workflow (ingest, diagnose, execute, reconcile). – Routed every bounded judgment (reply intent, playbook, tool choice) to Jev on the Decisions API as typed probabilities behind 4 model-free policy gates, so no model can move money; links sent via Razorpay MCP. – Engineered an asyncio worker pool over SQLite (WAL) processing 200 cases in 3.1s (p95 280ms) with an append-only audit trail and 443 passing tests; Dockerized and hosted on Microsoft Azure. RecallHQ – Self-Healing RAG | Python, LangGraph, pgvector, OpenAI, Streamlit 2026 – Built a RAG system over 2,200+ Slack messages: OpenAI text-embedding-3-small vectors in pgvector, fused with Postgres full-text search via RRF hybrid retrieval, orchestrated as a LangGraph StateGraph. – Designed a self-healing loop: the OpenAI Decisions API reranks candidates and gates evidence sufficiency; weak evidence triggers a query rewrite and re-retrieval, else it abstains; invalid citations are regenerated. – On 40+ labeled queries, hybrid retrieval with reranking lifted Recall@10 from 0.70 to 0.88 over dense-only with 95%+ citation precision; served via a Slack /recall bot and a Streamlit demo. DepthWizard – Single-Image 3D Height Mapping | PyTorch, ONNX, Hugging Face, Electron 2026 – Led a 6-member team on an ISRO problem statement at Smart India Hackathon 2026 to turn a single aerial or satellite image into estimated building heights and an interactive 3D scene. – Fine-tuned a pretrained DINOv3 vision-transformer encoder with a dense-prediction decoder on a remote NVIDIA H100 GPU using public remote-sensing height datasets, reaching 2.59 m RMSE with test-time augmentation. – Exported the model to ONNX, hosted inference on a Hugging Face Space, and shipped an offline Electron desktop app that exports results as GeoTIFF, GLB, and OBJ for GIS and 3D tools. Certifications & Achievements AIR 45 – ANRF AISEHack 2026, IIIT Hyderabad: Top 40 of 1,000+ teams in satellite flood mapping Oracle Certified – OCI Data Science Professional (2025) and Agentic AI Foundations Associate (2026) Google Cloud Study Jam – Top 20 (2026): Vertex AI and Gemini tracks; GDG On Campus, MGMU Open Source – PR #1277 merged into spicetify/spicetify-themes (23k+ stars); fixed a CSS layout regression Event Organizer – Swayambhu: Concept 2 Code (2025): designed problem sets and coordinated 100+ participants