Rajdeep Biswas
Data Scientist | Everything GenAI and LLMs
Bengaluru
5+ years of work experience
Jul 20242 yrs 3 mosPresent
• Designed and implemented prefix caching for a low-latency TTS inference pipeline, reducing repeated prompt computation across streaming requests and cutting average first-token latency by ~38% (from ~410ms to ~260ms).
• Built various Multi-Agent AI applications in the customer support chat space, using GCP Agent Development Kit, to boost containment of IVR calls by 55%, and service turnaround time by 30%.
• Introduced a novel LLM Confidence Measure mechanism using LLM API internal metrics to improve Generation Trustworthiness by 75%. A research publication is currently under the works in our research division of the company for this innovation.
• Developed an improved RAG System by implementing rejection mechanisms at multiple points in the standard pipeline to improve accuracy and reduce hallucinations by 25%.
• Finetuned custom lightweight language models and deployed them on GCP Model Registry as an endpoint, replacing vendor LLM intent classification and reducing costs by 90%.
Sep 20221 yr 11 mosJul 2024
• Developed multiple Prompt Engineering-based GenAI use cases, with Large Language Models (LLM) such as ChatGPT, Anthropic, Cohere, etc. tailoring solutions to diverse customer and developer personas.
Jun 20223 mosAug 2022
• Overlap analysis on textual data for a multiclass classification problem. Experiment on various preprocessing techniques to minimize class interaction and hence boost the accuracy of an AutoML provision from SAP (Data Attribute Recommendation).
Nov 20217 mosMay 2022
• Exploratory Data Analysis phase on F5 Load Balancer Telemetry data with HANA Infra Analytics.
• Conclusive feasibility check for an NL2SQL project. Assessing various pretrained models like GPT-J, T5 + WikiSQL, SQLNet, etc. on a limited training dataset.
Feb 20202 mosMar 2020
Kolkata, West Bengal, India
Mentored the 2018 (sophomore) batch of ~120 students on software development skills; had multiple contact classes on various technology, while still being a final year student at the same institution.
Jun 20193 mosAug 2019
Greater Bengaluru Area
• Took up various bugfixes on the frontend of the product, based on ReactJS. Responded to and fixed multiple internal and customer incidents and backlogs.
• Assessed the performance of various ML algorithms on matching intents and contributed to the NLP pipeline of a related product CoPilot.
• Contributed in migration of a few modules of the product's backend from Ruby to Java Spring Boot, the microservices framework.
Mar 20193 mosMay 2019
Kolkata, West Bengal, India
Developed a web application using Laravel that helped the company take in workshop attendees' information and register them, which was thus far carried out manually.
Mar 20193 mosMay 2019
Kolkata, West Bengal, India
Built a Flappy Bird clone using PhaserJS; only for educational purposes to be demonstrated to schoolchildren for TAL's early tech development and learning program to motivate the area of video game development and design.
Mar 20193 mosMay 2019
Kolkata, West Bengal, India
Created promotional material, with Adobe Premiere Pro CC 2017, for the company's workshops and seminars with multiple videos finalized for online publication.
Rajdeep’s AI Stack
Education
Master of Technology - MTech, Computer Software Engineering
Birla Institute of Technology and Science, Pilani
20242024
Bachelor's degree, Computer Science
Institute of Engineering & Management, Kolkata
20202020
Join Rajdeep and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal Boss