Manish Shaw

Manish Shaw -4 Year B.S. Male Economics Indian Institute of Technology Bombay Examination University Institute Year CPI/% Intermediate/+2 Matriculation CBSE ICSE TIGPS Hooghly Don Bosco Bandel - - Pursuing Minor in Artificial Intelligence & Data Science from CMiNDS, IITBombay Professional Experience AI Developer Intern— Tanumanasa • • • • [May’24-July’24] Worked as an AI Developer to make a Research-Paper oriented ChatBot to help with Comprehension Co-ordinate between the Team Members and delegate tasks to each person based on their Technical Prowess Actively Interacted with Clients by Pitching our Product, consistently gathering their Feedback and Insights to enhance Project Development, Improve User Experience, and ensure Client Satisfaction Devised a Plan of Action after taking their inputs and Work to improve the Efficiency and Reduce Costs Projects Undertaken Smart India Hackathon National Winners — ISL to Text and Vice-Versa [December’24] National Level Hackathon Conducted by Ministry of Education, Government of India • Selected for the Grand Finale from 7500+ Registered Teams and secured 1st Rank Nationwide • Designed and implemented LSTM and Multi-Headed Attention Architecture, achieving ISL recognition • Integrated Speech-to-Text, NMT and Mediapipe for rendering ISL gestures with 3D avatars for pipeline • Developed a Sign-to-Text system integrating Mediapipe’s Hand Tracking API, Pose Estimation, and Advanced NLP Models to convert ISL Gestures into Text, enhancing Accesbility for Non-Signing users. • Developed a System to detect Sign Language Users in Google Meet through UltraSonic Signals, ensuring their Active Speaker Visibility while ensuring Seamless Communication between Multiple Individuals iDEA Hackathon National Finalist — Transaction Network Fraud Detection [March’25] Hackathon Conducted by DoFS, Ministry of Finance Govt of India and Union Bank of India • Developed a Graph Construction Algorithm for fraud detection that creates Tripartite Network with an Artificial Fraud Node, incorporating Geographic proximity and Temporal Transaction Patterns • Optimized fraud detection systems by transforming discrete transaction data into d-dimensional continuous vector space representations leveraging Deepwalk-based representation learning on transaction graphs • Created Inductive Pooling to generalize embeddings for unseen transactions, boosting real-time detection. • Built a Stacked model combining Graph Neural Networks with XGBoost for Transaction risk assessment AI File Management System with Chatting Capabilities [Dec’24 - Jan’25] Hilti Hackathon Insititute Finalists conducted by EnPower, IITBombay • Developed a File Management ChatBot using RASA for Intent recognition and a Novel approach to Semantic searching of files featuring csv, images, audio, pdf based on Keywords Extracted from content • Parsed the semantic Vector Embeddings and used FAISS as VectorDB for Optimized Retrieval and Storage • ReRanked the Documents using FlashRank, thus most Useful Data is higher increasing Speed and Efficiency • Implemented Dynamic Memory Management for History-Aware responses, ensuring continuity, context Battery State of Health (SoH) Estimation [Oct’24 - Nov’24] igrenEnergi Hackathon Institute Finalists conducted by EnPower, IIT Bombay • Developed a Physics-Informed Neural Network model for Li-ion battery State of Health Estimation • Implemented advanced Data Preprocessing techniques, including Correlation Analysis to optimize Input • Achieved an Exceptional R2 score of 0.97 and a MSE of 1.79e-05, reflecting the Precision of the model Stock Prices Prediction using Time Series Data [May’24 - July’24] Summer of Code Project, WnCC IIT Bombay • Develop a Model for predicting stock market trends Leveraging State of the Art architecture like Long-Short Term Memory (LSTMs), Generative Adversarial Network (GAN) and Gated Reccurent Unit (GRU) • Use Historical Data to train models, capturing Complex Patterns and Dependencies in stock movements • The Model aims to enhance its ability to learn from Sequential and Time Series data, and generate more accurate predictions leveraging advanced algorithms for improved Performance, Reliability and Robustness Quadcopter Drone, MakerSpace [Jan’24 - Apr’24] Course Project, Guide Prof. Dinesh K Sharma • Collaborated in a team of 6 to successfully design, build, and test a fully functional Quadcopter Drone • Gained hands-on experience in Aerodynamics, Drone Mechanics, Electronics, and Teamwork • Contributed to Hardware Assembly, System Integration, and Software Control Enhancing Performance • Executed a 30-second test flight, during which the drone Flew Flawlessly under Professor’s Supervision Strategic Management Analysis of Lenskart [Oct’23 - Nov’23] Course Project, Guide Prof. Mayank Pareek • Analyzed Cost Leadership, Differentiation, and Organizational structure for effective insights • Conducted a Management Analysis, encompassing SWOT and TWOS assesments, to gain further insights • Strategically positioned the wide array of products and services of company in effective 3*3 GE Matrix Sentimental Analysis on Scraped Financial News [Jan’23] Self Project • Utilized Neural Language Processing to accurately understand and analyze market sentiments • Scraping Data from Web using libraries like Requests, CrawL4AI and Storing it in New Files • Developed the Neural Language Processing Algorithm using NLTK for efficient Text Analysis and Modeling • Imported a Special Dictionary curated for Financial Text in order to attain Maximum Accuracy Swiss Pairings and Leaderboard Updation [Feb’23] Self Project • Use Swiss Pairings (used in Chess Tournaments) to Effectively Pair Teams for Competition and rank them • Read Data using Pandas and Separate it into Groups using different Programming Languages for matches • Organize Matches between teams, ensuring each group competes separate from others for Fairness, Clarity Positions of Responsibility Manager, AI Community [Ongoing] Institute Technical Council, IIT Bombay • Lead a team of Junior Engineers on Institute Specific Projects and conduct Workshops and Events • Created a Flutter App for the Research-Paper Oriented ChatBot for Android and iOS using DART • Helped in the Development of INSTI-GPT which is a ChatBot trained on Institute Level Data • Participated in Hackathons collaborating with Cross-Functional Teams to develop Innovative Tech Solutions TechFest Co-ordinator—TechFest, IIT Bombay [Sep’24 - Feb’25] Selected from over 70+ applicants based on 3 tier selection criteria including peer reviews and an interview • Led Outreach Initiatives to Corporates, Securing Sponsorships and Partnerships contributing to budget • Collaboration plans with 10+ companies, driving Strategic Partnerships and maximized Sponsorship • Worked in a team of 50+ members involved in Publicizing events of the Techfest to 2500+ Colleges globally • Served as the Point of Contact (POC) for Multiple Companies during the Fest, ensuring Smooth Coordination, addressing their Requirements, and providing Exceptional support to enhance their Experience • Initiated a National Outreach Campaign across College networks, driving an increase in registrations for Techfest’s Web 3.0 workshop through Targeted Marketing and Student Ambassador collaborations Technical Skills Languages C/C++, Python, Git, Java, SQL, Bash, Fusion 360, LaserCAD Development HTML, CSS, LATEX, Docker, GitHub, Flutter, NLP, LLM, RAG, GraphRag Software & Packages NumPy, Pandas, Matplotlib, Seaborn, PyTorch, SciKit-Learn, TF, NLTK, Flask, FAISS, Hugging Face Transformers, LangChain, LlamaIndex, PyTorch, TensorFlow Courses and Certifications Docker — KodeKloud • • • • Understanding benefits of Containers in prod and Building custom Docker Images using Dockerfiles Persisting Data with Docker Volumes and Exposing ports to connect Containers with Host system Connecting Containers with each other using Docker Networking to Establish communication Managed Multi-Container applications using Docker Compose for Orchestration and Networking Relational Database Management System — PostgreSQL and BASH • • • Built BASH Script using Terminal and Scripting for Task automation, and building Interactive programs Utilized PostgreSQL for creating and managing Databases, Querying data, and Complex Data Analysis Utilized Git for tracking code changes, collaboration, and maintaining version control during this project