Help & Supportcareers@gravityer.com
Generative AI Engineer with 2+ years in GenAI systems, RAG pipelines, and FastAPI backends.
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Generative Al Engineer with production experience building GenAl systems, multi-agent orchestration platforms, RAG pipelines, and scalable FastAPI backends on AWS. Delivered systems handling 500+ daily queries at sub-200 ms p95 latency, reduced retrieval latency by 35%, and automated 15+ hrs/week of manual work. Deep hands-on expertise in LangChain, LangGraph, CrewAl, OpenAl API, Pinecone, FastAPI, Docker, and Python. Recognised with Rockstar of the Month and Pinnacle Award at Zepto. CGPA 9.6/10.
Bapurao Deshmukh College of Engineering, Wardha
B.Tech · Computer Science & Engineering
August 1, 2018 – June 30, 2022
Zepto
Generative AI Engineer
June 1, 2024 – Present
Mumbai, Maharashtra, India
Autonomous AI Employee System
June 23, 2026 – Present
• Architected multi-agent AI system: orchestrated 5+ autonomous agents for email automation, scheduling, and reporting using LangGraph + CrewAI — saving ~15 hrs/week of manual work. • Built persistent memory & integrations: implemented long-term episodic memory via Pinecone and connected Gmail, Slack, and Google Calendar via OAuth 2.0 and webhook-driven event handling. • Ensured reliability: added RAGAS evaluation suite with regression tests and failure-injection cases for non-deterministic agent testing.
View ProjectMultimodal RAG Knowledge Assistant
June 23, 2026 – Present
• Developed multimodal RAG system: unified ingestion and retrieval pipeline for PDFs, images, voice, and video (5+ formats) at sub-300 ms end-to-end query latency. • Engineered hybrid retrieval: OCR + Whisper STT + dense embeddings + re-ranking + metadata filtering for high-precision document QA. • Scaled async FastAPI backend with chunking optimisation and Pinecone namespace partitioning for multi-tenant knowledge base management.
View ProjectReal-Time AI Interview Copilot
June 23, 2026 – Present
• Built real-time AI assistant: WebSockets + streaming inference with sub-500 ms end-to-end latency, live Whisper speech-to-text, and streamed LLM output on React frontend. • Personalised contextual answer generation dynamically adapted to candidate's resume and target JD.
LLM Fine-Tuning & Evaluation Pipeline
June 23, 2026 – Present
• Fine-tuned Mistral-7B using QLoRA (4-bit quantisation) on a single GPU — achieved 18% improvement on task-specific benchmarks vs base model. • Built MLflow experiment tracking pipeline: hyperparameter logging, checkpoint management, model registry, and automated RAGAS eval (faithfulness, relevancy, context recall).
View ProjectAWS Certified Cloud Practitioner
Amazon Web Services
June 1, 2026 – Present
Generative AI Systems - LLMs, RAG & Prompt Engineering
Coursera / DeepLearning.AI
June 1, 2026 – Present
Spring Boot & Microservices
Udemy
June 1, 2026 – Present
Cultural Fit Analysis
The candidate demonstrates a strong cultural fit for a Generative AI Engineer role through diverse projects focusing on cutting-edge AI applications (multimodal RAG, real-time AI copilot, autonomous agents, LLM fine-tuning). Their experience at Zepto as a Generative AI Engineer directly aligns with the target role, showcasing practical application of advanced AI concepts in a production environment. The breadth of skills across AI/GenAI, ML/DL, Backend, and Cloud/DevOps, combined with open-source contributions, indicates a continuous learning mindset and a passion for the field.
Soft Skills & Operational Fit
The candidate's resume highlights significant achievements and leadership roles, such as 'Rockstar of the Month' and 'Pinnacle Award' at Zepto, indicating strong work ethic, impact-driven approach, and ability to deliver high-quality results. Leadership experience as 'Technical Head' and open-source contributions suggest initiative, collaboration, and a proactive learning attitude. The focus on performance metrics (latency, automation percentage) in project descriptions indicates a results-oriented mindset.