Jayakrishna Sahit Email & Phone Number

Bengaluru, Ka, In

Bengaluru, Karnataka, India

Bangalore Urban, Karnataka, India

[Visual Search @Flipkart]Architected and developed Flipkart's in house Visual search ML System. This includes the ingestion pipeline, retrieval pipeline as well as data pipelines for image logging. As a part of the ML system worked on multiple image models, vector matching engine (GCP ME) as well as implemented image processing algorithms such as de-duplication, category roll up and dominant object detectionDeveloped levers for latency, retrieval accuracy, cost, QPS[Capacity Forecasting in Supply Chain]Worked on the capacity forecasting data science problem in Flipkart's supply chain. Worked on introducing an ML model to replace the current statical approach being used[Model Monitoring]Working on creating an in house model monitoring framework.

Hyderabad, Telangana, India

Developed and tested algorithms for a new approach in camera to vehicle calibration, Li-DAR to vehicle calibration and camera intrinsic calibration for fisheye lens based on charuco/aruco targets. Developed their respective backend systems.Developed efficient sequence painting features using KDtrees to propagate labels across sequences, incorrectly, time synced or localityBuilt calibration profiles for users to be able to automate adding sensor profiles for calibrations

Implemented a data pipeline based on CycleGANS to help convert elastic seismic data to acoustic data for easier processing by XWI [Pytorch, AWS]Worked on improvements on the backend [Python]

Remote

Worked under DBpedia, in developing a tool for RDF extraction from textThe following is being done using various NLP related techniques- link to tool - http://text2rdf.linkeddata.es- link to work - https://github.com/sahitpj/TE-webapp

Bengaluru Area, India

Part of InMobi's Data Ingestion platform team - CDP (Customer Data Platform)Was involved in the migration of the platform’s schema service into a separate micro service entity

New Delhi Area, India

- Developed interactive content for the Visual Journalism Team in BBCTasked with Fake News research for the company. Investigated the whatsapp fake news stories which- Along side, was tasked with developing visual and interactive content as a part of it’s new initiatives as a part of visual journalism team.