Quickcall.dev

Founder (Dec 2025 - Present)

• Solving for developer productivity by solving for interruptions in their workflow
• Building QuickCall, an AI assistant that helps you connect your org tools and workflows without switching context

Yral

AI Consultant (Jan 2025 - Nov 2025)

• Built RAG based content moderation system (86.8% accuracy, +26%) using Phi-3.5 4B
• Deployed content moderation system on T4x2 GPU (19K req/USD, 3 req/sec)
• Optimized LLM inference with KV caching: ~70% cache hit rates and ~150 tokens/sec throughput on T4 GPU
• Developed real-time recommendation system with online learning capabilities, architected to support million-scale user interactions
• Architected end-to-end candidate generation and personalization pipeline with Airflow, GCP BigQuery, Redis caching
• Self-hosted MuseTalk talking head for video generation, supporting both image-to-video and video-to-video avatar synthesis

People+AI

AI Consultant (Nov 2024 - Dec 2024)

• Conducted systematic evaluation of 15+ existing Indic LLM benchmarks, mapping coverage gaps across 22 Indian languages
• Analyzed limitations in current evaluation frameworks: cultural context gaps, regional dialect coverage, and domain-specific assessments
• Researched ASR model evaluation approaches for Indian languages, identifying WER measurement inadequacies for multilingual contexts
• Designed comprehensive framework for India’s first standardized LLM leaderboard addressing identified evaluation gaps
• Authored strategic roadmap for 10 trillion token collection across Indian languages, published as a blog
• Educated NGOs on AI impact potential through STARS forum presentation, along with ideation sessions on AI for social good

smallstep.ai

Founder (Dec 2023 - Present)

Building Marathi LLM:

• Built custom SentencePiece tokenizer for Marathi, adding 15K tokens to address 3-5x token inefficiency in Llama’s English tokenizer
• Pretrained and Finetuneed Misal (Marathi LLM) 7B/1B models on 2B Marathi tokens using LoRA on A100 GPU
• Curated and cleaned 200K Marathi instruction dataset from Alpaca translations, along with IndicQuestionGeneration data
• Conducted systematic evaluation across reading comprehension, translation, sentiment analysis using internal QA dataset
• Achieved superior performance vs ChatGPT 3.5 on reading comprehension tasks, establishing first competitive Marathi LLM
• Open-sourced models, tokenizer, pretraining configs, and evaluation framework for regional language LLMs
• Published a technical blog with entire procedure here
• Launch : Linkedin Post, Twitter Post
• Coverage : YourStory, Moneycontrol, ARB Podcast
• Open Source : Huggingface Space

Medpiper

AI Consultant (Jan 2024 - Jun 2024)

Medical Documents Digitization:

• Led development of an AI platform for medical document digitization
• Engineered automatic annotation system for labelling and training
• Reduced processing time from 10 minutes to 1 minute, achieving a 10x performance boost

Tekion

Data Scientist (Aug 2022 - Jul 2023)

Document AI - Table Extraction:

• Developed a robust solution for extracting rows and columns from tables in images
• Trained a mask-RCNN based model to identify the structure of table components
• Reconstruction of table components for consumption

Automobile Service Recommendation:

• Conducted comprehensive EDA on seasonal trends in services taken
• Discovered key associations between different services taken
• Generated recommendations for individual vehicles based on their unique service histories.
• Impact : Significant increase in service add to cart and service taken rates

Pratilipi

Data Scientist (Dec 2020 - Aug 2022)

Modelling user item interactions using autoencoders:

• Leveraging bottle neck vectors as embeddings
• Creating a querying model to fetch similar interactions
• Built a collaborative filtering model
• Impact : High impact in reads and improvements in monetization observed

Creating hooks for increasing interaction at the end of each content:

• Leveraging Author similarity and common interactions of users
• Generating relevant author embeddings to capture category information
• Impact : 2x growth in reads after a user completed a read

Next authors to follow model:

• Applied clustering and probabilistic approach to compute “follow author” recommendations
• Impact : 20% increase in follow action from author profile page

Data Science Intern (Dec 2020 - Feb 2021)

Category personalisation “For you” section:

• Captured category interests and subcategory interests of users
• Personalised various category combinations
• Impact : Multi front impact on top line, monetisation and author follows observed

Conducted multiple experiments in “For you” section of the app:

• Tested multiple hypothesis to increase reads
• Conducted analysis to get insights of user behaviour across multiple geographies