Adversarial Search – Chess Bot
Dec 2024 – Jan 2025.
Designed a chess bot using Minimax with Alpha-Beta pruning for efficient move evaluation, reducing game-tree computation while preserving strategic quality and decision accuracy.
Experience & Projects
-
Jun 2025 – Present
Salesforce — Software Engineer (AMTS)
- Working on Revenue Cycle Management in Revenue Cloud with focus on billing and pricing infrastructure.
- Built a Billing Collection Agent on Agentforce and designed invocable actions to automate billing account summaries.
- Contributed to Salesforce Go by automating key workflows to improve process efficiency through intelligent task orchestration.
- Owned Revenue Cloud Growth license support in Salesforce Go, integrated with automation and provisioning flows.
-
May 2024 – Jul 2024
Samsung Research Institute, Bangalore — Software Developer Intern
- Collaborated with Samsung’s On-Device AI team to enhance quantization tools for optimizing AI models on devices.
- Developed features to extract encoding values and encoding files from ONNX and TFLite LLMs after quantization.
- Implemented AIMET-like encoding file generation for smooth compatibility with existing workflows.
- Conducted extensive validation to keep final model accuracy within 2% of the original floating-point model.
-
May 2023 – Jul 2023
Australian National University — Research Intern
- Worked on source-agnostic algorithms to minimize misinformation spread across large social networks.
- Implemented the Independent Cascade Diffusion Model to simulate misinformation propagation.
- Proposed an algorithm that improved containment factor by at least 60% compared to existing approaches.
- Results were published at the Pacific Rim International Conference on Artificial Intelligence (PRICAI), 2023.
-
2021 – 2025
IIT Delhi — B.Tech, Mathematics and Computing
- Graduated in May 2025 with B.Tech in Mathematics and Computing.
- Mathematics Society, IIT Delhi: Coordinator (Jun 2023 – May 2024) and Executive (Jun 2022 – May 2023).
- TRYST, IIT Delhi: Coordinator, Admin and Finance Vertical (Nov 2023 – May 2024), led finance model revamp and logistics for 20+ hybrid events.
- Hostel Technical Committee: Member (Jun 2023 – May 2024), managed hostel website and technical operations.
- Rendezvous, IIT Delhi: Activity Head (2023), managed hospitality and conducted 10+ events.
Projects
Virtual Memory Simulator – TLB Allocator
Course project with Prof. Ashutosh Rai, Aug 2024 – Sept 2024.
Built a simulator for FIFO/LIFO/LRU/Optimal TLB replacement and implemented a custom C memory allocator replicating malloc/calloc/free behavior using low-level system calls.
Dynamic CPU Process Scheduling Algorithms in C
Course project with Prof. Ashutosh Rai, Aug 2024 – Sept 2024.
Implemented FCFS, RR, MLFQ, and SJF schedulers with real-time input handling, and generated detailed metrics including burst, turnaround, waiting, and response times.
OpenHands
Open source contribution with Prof. G. Neubig (CMU), June 2024 – Aug 2024.
Integrated MemGPT-style short-term memory and built a dynamic context condenser that summarizes event streams near token limits, improving long-running agent stability and response quality.
Quantization of Large Language Models
Custom W8A16 quantizer for LLM optimization (May 2024).
Developed a custom PyTorch W8A16 quantization pipeline (8-bit weights, 16-bit activations), reducing memory footprint by 40%+ while maintaining practical inference performance.
OS Utility: Interactive Shell with Enhanced Features
Course project with Prof. Ashutosh Rai, July 2024 – Aug 2024.
Built a UNIX-style shell in C with built-ins, background execution, piping, environment variable support, and command history, along with robust error handling for interactive workflows.
Real-Time Disaster Recognition through Twitter Sentiment Analysis using BERT
Apr 2024 – May 2024.
Performed EDA and extensive tweet preprocessing (stopwords, URLs, symbols, emojis), then fine-tuned BERT for disaster tweet classification with validation accuracy above 87%.
Predictive Modeling for Student Grades
Project with Prof. Bhawani Sankar Panda, Mar 2024 – Apr 2024.
Built a student-grade prediction pipeline using TensorFlow Decision Forests (Random Forest and Gradient Boosted Trees), with feature engineering and hyperparameter tuning to achieve 85%+ accuracy.
Analysing Data to Identify Cyclone Features
Project with Prof. Samiran Mandal, Mar 2023.
Analyzed GHRSST and HYCOM ocean datasets to identify severe cyclonic storm indicators and thermocline/halocline structures, and used linear regression to observe rising sea-surface temperature trends.
Pattern Matching
Project with Prof. Ashish Chiplunkar and Prof. Naveen Garg, Oct 2022 – Nov 2022.
Implemented Rabin-Karp and modified KMP string matching (including wildcard support), achieving O(n + m) runtime over naive approaches and improved memory behavior via randomized prime-based hashing.
Network Data Transfer Capacity
Project with Prof. Ashish Chiplunkar and Prof. Naveen Garg, Oct 2022.
Implemented a modified Dijkstra + heap approach to compute maximum data transfer capacity in router networks, achieving optimized O(m log m) complexity where m is the number of links.
Search Nearby Program
Project with Prof. Ashish Chiplunkar and Prof. Naveen Garg, Sept 2022 – Oct 2022.
Built a range-tree based landmark database for map-style “search nearby” queries, enabling efficient spatial filtering with query handling in O(m + log² n) time.
Health Risk Prediction Model for Obesity and Cardiovascular Disease
Jan 2024 – Feb 2024.
Developed Random Forest and LightGBM models with feature engineering to predict obesity and cardiovascular risk from health and lifestyle factors, achieving around 91% accuracy.
ODI Winner Predictor
Personal project, Oct 2023 – Nov 2023.
Trained a Random Forest classifier on historical ODI cricket datasets to predict match outcomes using team-level and match-context features.
Particle Collision Simulator
Project with Prof. Ashish Chiplunkar and Prof. Naveen Garg, Aug 2022 – Sept 2022.
Designed a particle collision simulator with event-based processing and a custom priority queue to compute the first m collision events, including their times and coordinates, efficiently.
Python Interpreter / Advanced Python Interpreter
Project with Prof. Rahul Narain, Mar 2022.
Implemented an interpreter-style system to simulate constrained Python command execution, modeling core data representation and operation behavior for given inputs.