Dec 2025 – Present · 5 mos · Jaipur, Rajasthan, India · Hybrid

• Building SAKCHAM has been an incredibly rewarding journey. This platform was born out of a desire to make constitutional rights accessible to every Indian citizen through straightforward digital guidance. By integrating Generative AI into our design, SAKCHAM helps users navigate the often daunting process of documenting incidents like harassment or cyber fraud. In just four months since launch, partnerships with civil society groups have led to a remarkable increase in properly documented cases - over 50% more than before.
• The focus on user privacy is paramount; our design ensures secure storage and encrypted data export so users can document incidents safely. It’s not just about creating software; it’s about fostering a movement where individuals feel empowered and confident in pursuing justice. The goal is clear: transform complex legal processes into actionable steps that anyone can follow without feeling overwhelmed.
• Skills: Generative AI Solutions, User Experience Design, Legal Process Automation, Data Privacy Management, Community Engagement

Aug 2025 – Nov 2025 · 3 mos · Jaipur · On-site

• Built a pricing analytics model for 30+ aircrafts using structured inputs and cost breakdowns
• Supported decision-making by translating operational data into pricing insights
• Automated workflows and APIs to reduce manual work and improve turnaround time
• Worked cross-functionally to align product, operations, and data requirements

Apr 2025 – Aug 2025 · 4 mos · Lucknow, Uttar Pradesh, India · Hybrid

• Built end-to-end ETL pipelines processing 10,000+ resumes, reducing processing time by 40%
• Then add:
• Designed analytics-ready data structures for downstream reporting and insights
• Supported recruitment decisions through structured data and dashboards
• Used AI models as enablers, not the core value

Jun 2024 – Jul 2024 · 1 mo · Pune, Maharashtra, India · On-site

• Published two research papers: Runway FOD Detection and Crack Detection using advanced image processing techniques.
• Enhanced YOLO v8 for crack detection and YOLO v9 for FOD detection, improving accuracy and performance metrics for defense applications.
• Developed a GUI interface for an underwater detection system in a JRF-led defense project.
• Explored and contributed to a Ph.D. student project on self-driving cars, gaining insights into autonomous navigation and AI model integration.
• Collaborated with JRFs, professors, and Ph.D. scholars, gaining hands-on exposure to multiple defense research initiatives.

Feb 2023 – Present · 3 yrs 3 mos · Remote · Remote

• Key accomplishment includes :
• I build AI systems that replace manual processes, not just dashboards that describe them.
• I work with SMBs and startups to design, build, and deploy production AI - from computer vision pipelines on factory floors to LLM-powered chatbots that actually resolve queries. My systems don't sit in notebooks; they ship to real users and run in production.
• Key deliverables across 3+ clients:
• → Built computer vision workforce tracking (YOLOv9 + ArcFace) across 3 manufacturing units, 39 workers - cut manual payroll effort from 80% to 20%.
• → Deployed Gemini-powered chatbot + voicebot achieving 90% query resolution without human intervention, improving client retention by 13%.
• → Trained XGBoost demand prediction model (15+ features) that reduced raw material wastage by 56% for a footwear manufacturer.
• → Engineered 17-step automated ETL pipeline processing 8–15 GB per vendor from 8 sources into production-ready schema on AWS S3.
• What I build for clients: Custom ML models, LLM/RAG applications, computer vision systems, automated pipelines, AI chatbots, and end-to-end deployment (Docker, FastAPI, AWS).
• Skills: Python, PyTorch, LangChain, YOLOv5/v8/v9, FastAPI, Docker, AWS, SQL, XGBoost, RAG, Prompt Engineering, MLOps