Skip to content
View manastokale's full-sized avatar
🎮
Error 404: Life Not Found
🎮
Error 404: Life Not Found

Block or report manastokale

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
manastokale/README.MD
Obi-Wan Kenobi Hello There

Who Am I? 🔧

Typing SVG

Manas Tokale

I like building things.

I am a data scientist and machine learning engineer working across applied AI, forecasting, retrieval, and automation. I care about the business problem first, the model second, and shipping something useful at the end.

What I work on

  • LLM systems with retrieval, memory, and evaluation
  • Forecasting, risk, and decision-support models
  • Document AI pipelines and workflow automation
  • Product-minded data science that people can actually use

A few projects

  • Lineage: narrative AI workspace with time-aware retrieval and character memory
  • LocalPsych: privacy-first local mental health assistant fine-tuned with QLoRA
  • Synchrony Credit Intelligence: forecasting, fraud, segmentation, and credit recommendation engine
  • Railway Wheel Failure Prediction: predictive maintenance pipeline for 30-day-ahead failure detection

Outside of code

Cooking, badminton, painting, and video games.

Reach me

  • Email: manastokale@gmail.com
  • LinkedIn: https://www.linkedin.com/in/manas-tokale/
  • Portfolio: https://manastokale.github.io/portfolio/

✍️ Random Dev Quote

Pinned Loading

  1. Lineage Lineage Public

    A screenplay intelligence app that lets users explore TV scripts, query characters from a precise moment, and retrieve continuity-aware arcs and interactions.

    HTML

  2. LocalPsych LocalPsych Public

    Fine-tuned a pre-trained LLM for experiments in different domains like healthcare, psychology and customer support.

    Jupyter Notebook

  3. WorkforcePlanning WorkforcePlanning Public

    Hybrid forecasting system that predicts daily call volumes with XGBoost and allocates them across 48 half-hour intervals using learned intraday shapes

    Python

  4. gies-ai-experiments/VentureBot gies-ai-experiments/VentureBot Public

    repo to create multi agent simulations for teaching

    Python 5 6

  5. CreditRiskManagement CreditRiskManagement Public

    Four-part ML framework for spend forecasting, segmentation, fraud scoring, and credit limit recommendation across ~1.5M transactions and 18,070 accounts

    Jupyter Notebook

  6. RailwayWheelFailurePrediction RailwayWheelFailurePrediction Public

    30-day-ahead predictive maintenance pipeline using detector and mileage data to forecast wheel failure type in railway operations.

    Python