Kauntey Acharya

I am an AI Automation Engineer and Analyst at Innodata Inc., and I also work as an AI strategist and consultant. I help identify where AI can create practical leverage, translate business problems into AI solutions, and build working prototypes for real workflows: agentic workflows, RAG systems, research agents, API automations, and human-in-the-loop systems using n8n, Python, LLMs, and local AI.

Evaluation is a big part of my work. I test LLM behaviour, retrieval quality, workflow reliability, and automation performance to understand where and why AI systems succeed or fail. I bring the same quantitative toolkit (statistics, time-series modelling, and Bayesian inference) to quantitative analysis and finance.

Before industry, I was a pre-doctoral research fellow in theoretical physics at Ahmedabad University. My research focused on compact objects in general relativity, including naked singularities, black-hole mimickers, energy extraction, and high-energy particle collisions, and led to four peer-reviewed papers in journals such as Physical Review D and the European Physical Journal C.

Outside work I read research papers, play chess, and watch far too much sci-fi. I am also curious about neuroscience, cognitive science, and evolutionary biology.

1. AI Automation Engineer & Analyst

   Aug 2024 - Present

   Innodata Inc.

   Designing and building AI automations and workflows, alongside LLM evaluation, response-quality analysis, role-specific behaviour testing, and multimodal AI workflows.

2. Research Analyst and Writer

   Apr 2024 - Aug 2024

   NetBotPro Technologies

   Research-driven technical scripts, academic content review, SEO-aware writing, tracking recent research, and prompt engineering.

3. Pre-Doctoral Research Fellow

   May 2022 - Mar 2024

   Ahmedabad University

   Theoretical and numerical research on electromagnetic fields and particle dynamics in curved spacetime, with Prof. Pankaj S. Joshi and collaborators.

1. TickerBrief: Automated Equity Research Reports

   Type a ticker and get a full equity research report as an interactive web page and a PDF. An n8n workflow pulls prices, fundamentals, peers, SEC filings, and news in parallel; Python computes ratios, technical indicators, and sentiment; and a local LLM writes the narrative in three structured passes, with rule-based fallbacks.

   • n8n
   • Python
   • FastAPI
   • Ollama
   • yfinance
   • SEC EDGAR
   GitHub →Live demo →Note →

2. Model Monitoring + Drift Alert Agent

   A local MLOps stack that watches a deployed model. An n8n workflow scores new traffic, logs predictions to Postgres, and measures data, prediction, and target drift with Evidently AI. When a threshold is crossed, a local LLM explains the likely root cause in plain English and posts it to Slack, with a Grafana dashboard for the history.

   • MLOps
   • n8n
   • Evidently AI
   • Ollama
   • Postgres
   • Grafana
   GitHub →Live demo →Note →

3. ArXiv Research Analysis Agent

   Monitors arXiv every morning (gr-qc and q-fin by default), filters new papers by keyword, and sends a Telegram brief with a yes/no prompt. On “yes” it downloads the PDF, has a local LLM walk through the paper from first principles, and returns a Markdown report.

   • n8n
   • Ollama
   • ChromaDB
   • Telegram Bot API
   GitHub →Live demo →Note →

4. Local RAG Document Q&A Agent

   A fully local retrieval-augmented generation pipeline. Upload a PDF, TXT, DOCX, or MD file; it is chunked, embedded with nomic-embed-text, and stored in ChromaDB, and questions are answered by llama3.2 under a strict answer-only-from-context prompt. No cloud APIs, and no data leaves the machine.

   • RAG
   • n8n
   • Ollama
   • ChromaDB
   • Docker
   GitHub →Live demo →Note →

5. Banking & Financial Stock Analysis

   Quantitative analysis of banking stocks: returns, volatility, correlations, and risk metrics.

   • Python
   • yfinance
   • Pandas
   • Quant
   GitHub →

6. Gravitational-Wave Tests of GR (ppE formalism)

   Bayesian parameter estimation pipelines that test for deviations from General Relativity in gravitational-wave observations, using the parametrized post-Einsteinian (ppE) framework.

   • Parameter estimation
   • Bayesian inference
   • Python
   • Gravitational waves
   GitHub →

7. Credit Score Data Analysis

   EDA of credit datasets focused on default risk, class imbalance, correlations, and interpretability.

   • EDA
   • Pandas
   • Risk
   GitHub →

8. Indian Startup Investment EDA

   Exploratory analysis of Indian startup funding trends, sectors, and investor activity.

   • EDA
   • Pandas
   • Visualisation
   GitHub →

1. JMN-1 singularity in weak magnetic field

   M. Azreg-Ainou, K. Acharya, P. S. Joshi

   Eur. Phys. J. C 84, 535

2. High energy particle collision in vicinity of naked singularity

   K. Acharya, P. Bambhaniya, P. S. Joshi, K. Pandey, V. Patel

   Phys. Dark Univ. Vol. 50, 102101

3. Energy extraction from Janis-Newman-Winicour naked singularity

   V. Patel, K. Acharya, P. Bambhaniya, P. S. Joshi

   Phys. Rev. D 107, 064036

4. Rotational energy extraction from Kerr black hole’s mimickers

   V. Patel, K. Acharya, P. Bambhaniya, P. S. Joshi

   Universe 2022, 8(11), 571

• Loading…

All notes →

Ping me for opportunities related to AI roles, agent builds, research collabs, or just a good paper to discuss.