system online · dublin, ireland · stamp 2 visa

ANKUSH DERLE

~/ai$

I architect production-grade AI systems — autonomous agents, RAG pipelines and multi-LLM orchestration — and ship them end-to-end: FastAPI backends, vector search, Docker fleets and CI/CD.

0
Years Gen-AI
0
Docs in RAG
0
Live microservices
0
Uptime
⬡ AGENT_CORE :: ACTIVE
RAG · LLM-OPS · MCP
Ankush Derle
OPEN TO OPPORTUNITIES
01 profile.load()

Intelligence that
ships to production

I build AI systems that survive contact with the real world. At Adeptia I architected LLM-directed platforms for data mapping and knowledge retrieval — RAG chatbots over 10,000+ enterprise documents, rule engines processing 50,000+ records a day, and FastAPI services answering in under 200ms at peak load.

Now I'm building Vignova — a full-stack AI career platform with autonomous job-application agents and a hybrid orchestration layer across three LLM providers — while pursuing my MSc in Artificial Intelligence at the National College of Ireland, Dublin.

My obsession: the gap between an impressive demo and a system you can trust — cost routing, hallucination control, observability and uptime.

ankush@dublin: ~/profile
0
Manual effort cut
0
Retrieval relevance ↑
0
Hallucinations ↓
0
Records / day
02 experience.log

Production AI,
measured in outcomes

Software Engineer — Gen AI

ADEPTIA · INDIA
APR 2023 — JUL 2025
AIMap — AI Data Mapping Platform RAG Chatbot Python Rule Engine
03 flagship.exec()

Vignova.io — an
autonomous career platform

Job post → submitted application
in under 2 minutes.

A full-stack AI SaaS (Next.js + FastAPI + PostgreSQL + Docker) architected and shipped solo — resume generation, job tracking and fully automated applications across 4 core modules and 6 containerised services.

JAN 2026 — PRESENT
AGENTIC WORKFLOW — LIVE
👁️
Observer
reads the job page
🧠
Planner
maps the steps
Executor
applies autonomously
AVG APPLICATION TIME: 20 MIN → <2 MIN
🧬

GenAI resume engine

DeepSeek (Ollama), Gemini & Sarvam with intelligent model routing — API costs down 45%.

🔀

Hybrid LLM orchestration

Cost / latency / quality routing across 3 providers; RAG lifted ATS match scores 30% in A/B tests.

🧩

Chrome Extension (MV3)

Job scraping, one-click resume tailoring and auto-apply across major job portals.

📦

6 services · 99.5% uptime

Docker Compose fleet behind an Nginx reverse proxy, sustained in production.

🚀

CI/CD automation

GitHub Actions pipelines cut release cycles from 30 minutes to under 5.

🗂️

4 core modules

Resume generation, job tracking, tailoring and automated applications — one platform.

04 stack.scan()

The toolkit behind
the systems

Generative AI & LLMs

RAGLangChainLangGraphAgentic WorkflowsMulti-Agent SystemsMCP ServersPrompt EngineeringFine-TuningVector EmbeddingsNLP

Models

GPT-4ClaudeGeminiLLaMADeepSeekGrokSarvam

ML & Data Science

RegressionDecision TreesModel EvaluationFeature EngineeringEDAData Visualization

Languages & Frameworks

PythonFastAPIPyTorchTensorFlowKerasScikit-LearnPySparkNumPyPandasREST APIs

Data & Backend

MicroservicesAsync ProcessingETL PipelinesPostgreSQLMySQLMongoDBMilvusFAISSPinecone

Cloud, DevOps & AI Tooling

DockerAWSAzureGitHub ActionsNginxModel DeploymentCursorClaude CodeCopilot
05 education.read()

Foundations &
frontier

PURSUING
MSc in Artificial Intelligence
National College of Ireland · Dublin
2021
Bachelor of Engineering
Savitribai Phule Pune University · India
06 contact.init()

Let's build something
intelligent together.

Open to AI Engineer, ML Engineer and Gen-AI roles in Dublin and across Ireland — on-site, hybrid or remote.