Help & Supportcareers@gravityer.com
AI Engineer with 6+ years in LLM-powered Applications & Agentic AI Workflows
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI/ML Engineer with 6+ years of experience designing and deploying intelligent systems across healthcare, fintech, and recruitment domains. Specialises in LLM-powered application development, agentic Al workflows, RAG pipelines, and multi-agent orchestration using LangChain, LangGraph, and OpenAI/Anthropic APIs. Proven ability to translate complex business problems into production-ready Al systems - from rapid prototyping to scalable cloud deployment on AWS. Passionate about leveraging cutting-edge Al to automate workflows, surface insights, and deliver measurable impact in real-world environments.
International Institute of Information Technology, Hyderabad
B.Tech · Computer Science
August 1, 2016 – June 30, 2020
Freelance AI Engineer
Freelance AI Engineer
June 1, 2025 – May 31, 2026
Hyderābād, Telangana, India
Saven Technologies
Software Engineer
December 1, 2023 – May 31, 2025
Hyderābād, Telangana, India
Sids Farm
Freelance Software Engineer
February 1, 2021 – November 30, 2023
India
Cure.fit
Software Engineer
June 1, 2020 – January 31, 2021
Bengaluru, Karnataka, India
Cure.fit
Software Engineer Intern
May 1, 2019 – July 31, 2019
Bengaluru, Karnataka, India
Document Q&A System over PDFs
June 24, 2026 – Present
Built an end-to-end document Q&A system ingesting PDFs and docs, embedding content into Pinecone, and answering natural language queries with source attribution via a FastAPI REST endpoint. Implemented hybrid search, re-ranking, and metadata filtering; evaluated with RAGAS for context precision, recall, and answer faithfulness in production.
Real-Time Speech-to-Text + Summarization Pipeline
June 24, 2026 – Present
Built a real-time audio pipeline streaming microphone input over WebSockets, transcribing with Whisper, and generating LLM summaries end-to-end latency under 3 seconds. Handled chunked audio streaming, speaker turn detection, and incremental summarisation for meetings, lectures, and interviews.
Autonomous Agent with Tool Use
June 24, 2026 – Present
Built a fully autonomous LLM agent dynamically selecting and invoking tools (web search, code execution, external APIs) to complete multi-step tasks using ReAct-style reasoning loops with error recovery. Exposed agent capabilities via a FastAPI streaming endpoint; implemented structured tool outputs and graceful failure handling for production reliability.
Multi-Agent Collaboration System
June 24, 2026 – Present
Architected a multi-agent system with specialised sub-agents (researcher, planner, coder, reviewer) orchestrated by a supervisor — addressing agent loops, context overflow, and conflicting outputs. Benchmarked against single-agent approaches, demonstrating measurable improvements in task completion and output quality for complex research and code generation tasks.
LLM Evaluation Framework
June 24, 2026 – Present
Built a custom evaluation framework to benchmark LLMs (GPT-4, Mistral, Llama 3) on domain-specific tasks measuring accuracy, hallucination rate, latency, and cost across standardised test sets. Integrated RAGAS for RAG-specific metrics and LangSmith for end-to-end tracing; surfaced results in a structured dashboard for data-driven model selection.
Fine-Tuning Pipeline with LoRA/QLoRA
June 24, 2026 – Present
Built a reusable fine-tuning pipeline for instruction-tuning LLMs on domain-specific datasets using QLoRA — 4x GPU memory reduction vs full fine-tuning while retaining model quality. Covered full pipeline: data cleaning, instruction formatting, supervised fine-tuning with PEFT/LoRA, checkpoint management, and evaluation against held-out test sets.
RAG Pipeline with Evaluation & Telemetry
June 24, 2026 – Present
Built a production-grade RAG pipeline with full observability — OpenTelemetry/LangSmith tracing for retrieval quality, latency, and answer faithfulness monitoring. Implemented RAGAS automated evaluation measuring context precision, recall, and answer relevancy to iteratively improve pipeline quality.
MCP Server & Offline Privacy-First AI App
June 24, 2026 – Present
Designed a custom MCP server exposing domain-specific tools (file I/O, DB queries, API calls) to Claude for structured tool use in agentic workflows. Built a fully local AI application using Ollama + Llama 3 with on-device RAG (nomic-embed-text + ChromaDB) zero data leaves the machine, viable for healthcare, legal, and enterprise compliance use cases.
Full-Stack Web App - Reddit Clone
June 24, 2026 – Present
Designed and implemented RESTful APIs with FastAPI; API patterns directly applicable to building LLM inference and embedding endpoints for GenAI products.
Distributed File-Sharing System - Mini Torrent
June 24, 2026 – Present
Built a P2P file-sharing system with parallel multi-piece downloads demonstrates concurrency and systems-level Python programming skills.
Cultural Fit Analysis
The candidate's project diversity, spanning document Q&A, autonomous agents, real-time speech processing, multi-agent systems, and LLM evaluation, indicates a broad interest in AI applications. Their experience in healthcare, fintech, and recruitment domains shows adaptability and a willingness to tackle varied challenges. The focus on building production-ready, scalable, and privacy-first solutions aligns with a culture that values innovation, reliability, and ethical AI development. The freelance experience also suggests self-motivation and initiative.
Soft Skills & Operational Fit
The candidate's resume highlights full system ownership, cross-functional collaboration, and technical documentation, indicating a strong operational fit. The project descriptions demonstrate a proactive approach to problem-solving and a focus on delivering measurable impact, which aligns well with senior engineering roles. The diverse project portfolio suggests adaptability and a strong learning curve.