Help & Supportcareers@gravityer.com
AI Engineer with 1+ years in LLM, RAG & Agentic Systems
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 building production LLM, RAG, and agentic systems, with a live AI product (TryOwn) serving real users and a deep-learning internship in computer vision. Backend experience at Orange Business on blockchain-based cloud storage. Comfortable across Python, LangChain, FastAPI, vector search, multi-agent orchestration, and full-stack development.
Netaji Subhas University of Technology
B.Tech. · Electronics & Internet of Things (IoT)
August 1, 2020 – June 30, 2024
ORANGE BUSINESS
Backend Engineer (Account Associate)
June 1, 2025 – Present
Gurgaon, Haryana, India
EIGENGRAM
AI Engineer Intern
November 1, 2024 – May 31, 2025
India
TRYOWN (Live Virtual Try-On Fashion Platform)
June 22, 2026 – Present
• Built and shipped a full-stack AI fashion platform live with real users, featuring virtual try-on, pose transfer, a marketplace, and a community forum • Engineered a recommendation engine using FashionSigLIP 512-d embeddings in pgvector, with interaction-weighted vector blending for cosine-similarity ranked feeds • Served model inference through a FastAPI service; built forum feed-ranking (hot, rising, top, new) with nested comment threading and Redis-based per-user rate limiting
Autonomous macOS AI Agent
June 22, 2026 – Present
• Designed a 7-tier cost-tiered execution engine (shell, URL schemes, AppleScript, Chrome JS, Accessibility API, vision-LLM, GUI), auto-routing tasks to the cheapest working route and cutting average token cost ~10× vs. a pure vision loop • Built context engineering for long sessions: a 3-pass history curator (screenshot collapse, failure deduplication, LLM compaction) keeping the context window clean across 14-iteration runs; episodic memory with Bayesian-smoothed per-tier priors and fcntl.flock safe concurrent writes • Built a harness-based evaluation system: frozen regression corpus of 10 cases, keyword + LLM-as-judge scoring, predicate-based offline evaluation, and pytest integration for CI • Designed a multi-agent orchestration layer with 3 topologies (N-way debate; Planner → Researcher → Executor → Verifier pipeline; parallel fan-out), a Planner-Critic-Judge flow that adversarially reviews plans before any side-effecting action, plus a shared blackboard, asyncio message bus, and resource locks on keyboard/mouse/chrome_tab
View ProjectPDF RAG Application
June 22, 2026 – Present
• Built a retrieval-augmented generation app for Q&A over PDFs, docs, and URLs using semantic chunking and FAISS vector search • Designed prompt templates and a retrieval-evaluation loop to reduce off-topic and hallucinated answers
OCR Text Extraction Tool
June 22, 2026 – Present
• Built a batch document pipeline with preprocessing (deskew, denoise, contrast enhancement) to improve extraction on noisy scans, automating conversion to structured JSON with per-field confidence scores, tested across receipts to academic papers
Cultural Fit Analysis
The candidate's diverse project portfolio, ranging from AI fashion platforms to autonomous agents and RAG applications, indicates a broad interest in AI/ML domains and a proactive learning attitude. The experience with both AI engineering and backend development, including blockchain, suggests adaptability and a willingness to explore different technical areas. The personal projects demonstrate initiative and self-direction, which are valuable traits for cultural fit in innovative environments. The target role of 'AI Engineer' aligns well with the candidate's demonstrated skills and project focus.
Soft Skills & Operational Fit
The candidate demonstrates strong problem-solving skills through complex project designs (e.g., 7-tier execution engine, 3-pass history curator). The project descriptions indicate an ability to work on end-to-end solutions, from model engineering to deployment and user interaction. The experience with competitive programming suggests a disciplined and analytical approach to problem-solving. The candidate's ability to ship a live product (TRYOWN) indicates a results-oriented mindset and operational awareness.