WASD move · Shift boost · Space drift · E enter · F walk · H horn · M map · P photo
0:00.0Esc gives up
🚨 WANTED
the long arm of the law
BUSTED.
The cops caught up with you. The city forgets everything you did: coins, skills, quests, stickers, the lot. Drive nicely this time?
Photo modeWASD pan · scroll or +/− zoom · Esc exitsDrag the stick to pan
boost
gps
City map
Tap the map or pick a place to set a route. M or Esc closes.
play mode
Welcome to the neon city.
Every building is part of my portfolio. Drive up to a glowing door and step inside. Start at the plaza right in front of you, where your motorcycle is parked too.
W A S D or arrows to drive
E or Enter to go inside
F to hop out and walk
Shift boost · Space handbrake drift · H horn
On a phone: drag the stick, tap the door button, tap Walk to hop out
M for the map and GPS · P photo mode
Tap the minimap for the full map and GPS
Grab ⚡ skill orbs to find out which role you'd suit. Drive carefully: hit someone and the cops come.
Open ◎ Missions for side missions
side missions
Missions
Progress lasts until you refresh the page. Every visit starts a fresh game.
Know the city
Step inside every building.
Skill shards
Shards are hidden in alleys and behind buildings. Hop out of the car (F) and find them on foot.
Arcade champion
Earn a gold medal in Career Run at the Arcade.
Side quests
People with a ! over their heads need a hand.
Sticker book
Twelve stickers are hidden around town, in corners, alleys and odd spots.
Achievements
You have 0
Coins come from the road, drifts and helping people.
skill hunt
Your skills
spawn point · read me
Welcome to Adhith Sankar's city
A GenAI engineer with 2+ years of production experience building LLM, RAG and multi-agent systems for enterprise clients. He takes products from whiteboard to deployment, and measures whether they actually work.
2+ years taking GenAI products from whiteboard to deployment at Entelegen Analytics LLP: architecture, backend APIs, containerised deployment on SAP AI Core, GCP and AWS, and client handover.
RAG pipelines, LangGraph multi-agent orchestration, and turning vague business problems into production AI systems people actually use.
Evaluation. Shipping an LLM system is the easy half; knowing whether the last prompt or model change made it worse is the hard half. That is what I build now.
Data Scientist · Entelegen Analytics LLPAug 2024 – PresentCoimbatore, India
Esc or × to get back in the car
AGENT · handed over to production
Agentic Purchase Requisition Automation
LangGraph multi-agent system that automates purchase requisition workflows: each agent owns one step, the orchestration layer manages state and handoffs, and new agents plug in without changing the core graph.
Flask + WebSocket backend for a real-time chat interface, plus a parallel n8n low-code version the client can modify without engineering support.
Ran client demos and iterated on agent logic and UX through to production handover.
LangGraphLangChainFlaskWebSocketn8nDocker
Esc or × to get back in the car
RAG · 100+ contracts processed
Contract Document Intelligence: RAG Q&A
RAG pipeline over enterprise contract PDFs: text, tables and images extracted, embedded with Sentence Transformers and indexed in FAISS on SAP Blob Storage.
A locally hosted Mistral model served via Ollama, so sensitive contract data stayed inside the client's infrastructure.
Ingestion and inference run as separate Dockerised services on SAP AI Core; used by the client team for contract review.
FastAPIFAISSSentence TransformersOllama · MistralSAP AI CoreDocker
Esc or × to get back in the car
RETRIEVAL · live on a customer-facing site
UNSPSC Commodity Recommendation Engine
Maps free-text purchase descriptions to UNSPSC commodity codes, replacing a manual classification step: cosine-similarity retrieval over the embedded taxonomy, with a Mistral layer for ambiguous inputs.
Shipped as a Dockerised Flask REST API on SAP AI Core, then migrated to AWS Fargate as usage grew; live on the client's procurement site across multiple organisations.
FlaskEmbeddingsCosine similarityMistralSAP AI Core → AWS Fargate
Esc or × to get back in the car
RAG · deployed for a Chennai school
NCERT Knowledge Assistant
Contributed to a RAG learning assistant over NCERT textbooks, deployed for a leading Chennai school.
Built the ingestion pipeline (text and image extraction, images to Cloud Storage, embeddings with metadata in Qdrant) and FastAPI/WebSocket services for chat Q&A with image citations and a quiz mode.
FastAPIQdrantGemini APIMongoDBKnowledge graphsGCP Cloud Run
Esc or × to get back in the car
flagship project · evaluation
RAG Hallucination Cascade
Catching made-up answers from a RAG system without paying an LLM judge to check every answer: free rules first, a small trained classifier next, and the judge only for what the first two can't settle.
35%of answers sent to the judge
92%of hallucinations caught (11 of 12)
$2.60per 1,000 answers
0.19 stypical check time
The classifier didn't beat a length rule: Test ROC-AUC 0.71 vs 0.80 for answer length, intervals overlapping. Its value is cost: about 10 points fewer judge calls at the same recall.
Hallucinations looked more like the contract, not less: They reuse its wording while misreading it, so word overlap was useless (AUC 0.52). Embeddings measure topic, not truth.
The judge was checked against people: 60 answers labelled blind: 67% agreement (κ 0.25) at first, mostly from inconsistent refusal labels. Reported as is, not hidden.
The guardrail catches phrasings, not intent: 19 of 24 test attacks blocked; all 5 reworded attacks written afterwards got through.
40 held-out answers, 12 of them hallucinations, with thresholds chosen on training data first. 95% intervals: 20–50% sent to the judge, 71–100% caught. *Partly by construction: the test reference uses the judge's label where the human labeller agreed with it.
PythonFAISSQwen3 embeddingsMistral 7B (local)scikit-learnClaude as judgeStreamlitCUAD dataset
B.Tech, Computer Science and Engineering — SRM Institute of Science and Technology · 2019–2023 · CGPA 8.9/10
Coronary Artery Disease Prediction using ML Algorithms — IEEE ICESIC 2022
Machine Learning Intern, TechAnalogy (Jun 2021 – Feb 2022): Integrated NLP/ML models into web services; built DataCLI, a Python (Click) CLI that automates preprocessing.
Data Associate Intern, ExpandAI (Aug 2021 – Nov 2021): Annotated training data and ran object-detection model experiments.
Student Research Assistant, SRM University (Nov 2021 – 2022): Co-authored 'Coronary Artery Disease Prediction using ML Algorithms' (IEEE ICESIC 2022).
Esc or × to get back in the car
side project
Self-hosted n8n automation server
A personal automation server in Docker with persistent workflows: the same low-code tool I shipped to a client as an alternative to a hand-built agent.
n8nDocker Compose
Esc or × to get back in the car
certification · stack
Cloud tower
Google Cloud Professional Data Engineer (Aug 2026 – Aug 2028)
DockerSAP AI CoreGCP Cloud RunCloud StorageAWS Fargate
Esc or × to get back in the car
arcade · side game
Career Run
Run from SRM in 2019 to today. Jump the bugs, rate limits and hallucinations, collect the skills, and read each chapter as you pass its gate. About 45 seconds.
Esc or × to get back in the car
post office · contact
Send me a message
Open to GenAI, LLM and AI Engineer roles. Open to relocating anywhere in South India.
You've seen every building, found every skill shard and taken gold in Career Run. That's more attention than most CVs get, so thank you. If you're hiring for GenAI, LLM or AI Engineer roles, I'd love to hear from you.
Jump the bugs, rate limits and hallucinations, but stay low under drones. Collect skills in a row for a combo. Every gate is a chapter of my career, and bumping into things only slows you down.
Space / ↑ / click to jump, hold to jump higher
Tap anywhere to jump, hold to jump higher
Double jump unlocks when you reach your first full-time job
you made it to now
Career complete.
Chapter recap: the run as a timelineSkills you collected