AI Engineering
Generative AI, multi-LLM orchestration, structured output validation, RAG pipelines, fine-tuning, and systematic evaluation for dependable production workflows.
Aniket Thakur · AI × Product × Finance
Building AI systems, products, and intelligent workflows.
I’m Aniket Thakur — an AI researcher and product builder working across Generative AI, machine learning, fintech, and data-driven products.
Aniket Thakur · New Delhi / Okinawa
APM & AI Researcher
3.82/ 4.0
GPA · Rank 2
B.Sc. Data Analytics, DSEU
OIST
AI Research Intern
Okinawa, Japan · Graph ML & Multi-LLM
70%
Cycle Reduction
Laya Automated SEO Intelligence System
APM
Drivio Technologies
Dec 2025 – Present · Loan journeys & LMS
Focus & Expertise
Three areas I work across.
Technical rigor combined with product ownership and financial context. Building systems where machine learning models drive measurable business decisions.
Product Ownership
Leading end-to-end loan journeys in two-wheeler and vehicle marketplaces, integrating core banking APIs for KYC, disbursement, and repayment in the LMS, and managing HRX/HRMS operations.
Financial Engineering
Pursuing an M.Sc. in Financial Engineering at WorldQuant University. Applying quantitative analysis, risk mechanics, and algorithmic credit automation to real-world financial architectures.
Selected Work
Featured systems & case studies.
Featured systems designed to automate workflows, solve data ambiguity, and deliver quantifiable operational outcomes.
01 / Flagship System
Laya — Automated SEO Intelligence & Generation System
An automated SEO pipeline that connects live search evidence directly to structured content generation, eliminating manual friction in content production.
Web Scraping
SERP Analysis
Multi-LLMs
Schema Validation
70%
Reduction in content-creation cycle turnaround time
02
Data Product
Prompt-Based Business Intelligence System
Natural-language query engine over SQL databases enabling non-technical stakeholders to perform self-serve exploratory analysis and executive reporting.
Natural Language
SQL
Data Analytics
03
Workflow Automation
osTicket — AI / Automated Ticketing Workflow
Integrated LLM categorization and automated response routing into support desk operations to triage tickets by urgency and reduce resolution lag.
LLM Routing
Triage Automation
Support Desk
04
Generative AI
AI Content Generation & Multi-LLM System
Multi-model orchestration framework utilizing structured output validation to deliver consistent, deterministic copy across multi-channel marketing campaigns.
Multi-Model
Structured Outputs
Orchestration
05
EV Infrastructure
FastCharger — EV Charging Discovery Infrastructure
A mobile-first EV charging discovery platform foundation with provider ingestion, data normalization, geospatial PostgreSQL/PostGIS storage, typed APIs, and India-first routing.
Next.js 16
TypeScript
PostgreSQL
PostGIS
Drizzle ORM
Zod
Vitest
Track Record
Career progression.
Experience across high-growth fintech, international AI research, and early-stage startup advisory.
Gurugram, India
Associate Product Manager
Drivio TechnologiesOwn end-to-end loan journeys for a two-wheeler marketplace, align cross-functional stakeholders, and integrate core banking APIs for KYC, loan disbursement, and repayment in the Loan Management System (LMS). Managed HRX/HRMS operations and built LLM content-generation agents with structured output validation, optimizing latency and consistency.
Loan Journeys
Core Banking APIs
LMS & HRX
LLM Validation
Okinawa, Japan
AI Research Intern
Okinawa Institute of Science and Technology (OIST)Developed Graph Neural Network (GNN) algorithms to analyze complex ecological food webs and model energy-flow dynamics, boosting predictive accuracy. Engineered data pipelines using multi-LLM ensembles (Gemini, DeepSeek, Anthropic, OpenAI) to synthesize semantic representations from unstructured scientific literature.
Graph Neural Networks
GraphSAGE & GAT
Multi-LLM Ensembles
Complex Systems
Gurugram, India
Business Analyst
Drivio TechnologiesLed AI implementation for a used-car finance system, deploying LLM automation to accelerate loan processing and automated credit decisions. Fine-tuned marketing LLM content agents, automating multi-channel asset creation across acquisition channels.
Credit Decisions
Fintech Automation
Fine-Tuning
Marketing AI
New Delhi, India
Business Consultant & Analyst
Analy AssistFacilitated go-to-market (GTM) launches for four early-stage startups across food, health, and fashion through market research and cross-functional execution. Delivered rigorous cost analysis and executive reports that reduced operational expenses by 15%.
GTM Launches
Cost Analysis (-15%)
Startup Advisory
Academic Research · OIST Japan
Graph Neural Networks for ecological food-web modelling.
At the Okinawa Institute of Science and Technology (OIST), my research focused on computational modeling of complex ecological networks. Food webs describe intricate trophic dependencies and energy-flow dynamics across species communities.
By developing inductive Graph Neural Network algorithms (incorporating GraphSAGE and Graph Attention Networks), we modeled non-linear trophic interaction probabilities, capturing topological centrality and community structures that traditional pairwise models overlooked.
Alongside structural GNN modeling, I engineered semantic synthesis pipelines leveraging multi-LLM ensembles (Gemini, DeepSeek, Anthropic, and OpenAI) to standardize, cross-verify, and ingest ecological literature into structured topological datasets.
Academic Background
Education & foundations.
Mathematical, statistical, and quantitative groundwork in financial engineering and data analytics.
M.Sc. in Financial Engineering
WorldQuant UniversityAdvanced quantitative modeling, portfolio theory, derivative pricing, and algorithmic risk management.
B.Sc. in Data Analytics
DSEU Shakarpur-2 Campus, New DelhiStatistical computing, predictive modeling, econometrics, and database engineering.
From the blog
Latest Articles.
Technology, Product, Auto, and Finance.
Agentic Coding: Can AI Build, Test, and Deploy an Entire Application?
Compare inline AI autocomplete with autonomous coding agents in 2026. Learn how terminal agents build, test, debug, and open PRs safely under CI/CD controls.
AI Agent Development Skills: Tools, Memory, and Orchestration
Learn practical AI agent development skills in 2026 using Pydantic AI and LangGraph: typed tool schemas, dependency injection, state reducers, and loop bounds.
AI App Builders vs Traditional Development: What Should Startups Use?
Compare AI app builders ('vibe coding') with traditional software engineering Using Next.js and Payload/Strapi in 2026: speed, technical debt, security, and ownership.