Summary
GenAI practitioner who shapes AI strategy and builds the systems underneath it — from AI Over BI natural-language analytics and finance data-transformation PRDs to production LLM backends, ML/CV pipelines, and low-level GPU/Rust/WASM systems shipped end to end. Backed by an MBA in Business Analytics and hands-on delivery on AWS and GCP.
Experience
Full Stack LLM Development Analyst — Finance Data Transformation — Accenture
Bangalore, India | Apr 2026
- Built AI Over BI natural-language analytics on an in-house framework — grounded the system in finance data models via semantic-layer mapping, tested that NL queries resolved to the correct metrics and dimensions, and automated JSON configs (Google Apps Script) for 13+ usecases.
- Owned requirements-to-acceptance with business and strategy stakeholders for 3+ finance data transformation usecases — translated needs into source-to-target specs and Claude-assisted PRDs, drove BRD/PRD sign-off, and drove UAT discussions to validate delivered outputs.
- Owned the data layer testing framework — a custom SQL test harness running source-to-target reconciliation alongside null, duplicate, and referential integrity checks across finance data transformation usecases, automating cross-system checks and reducing manual QA effort.
- Took ownership of RBAC across 5+ systems (GCP projects, Git repositories, internal tools) without documentation — mapped the permission model, defined role tiers, executed assignments, and established a shared Git repository with tiered approval workflows.
Full Stack LLM Development Analyst — GenAI Experience Platform — Accenture
Bangalore, India | Jan 2026 – Mar 2026
- Owned the design and deployment of a production-ready backend pipeline enabling end-to-end image generation, persona detection, and secure UI workflows using AWS Lambda, API Gateway, and containerized services, validated through full E2E testing.
- Implemented enterprise-grade integrations and resilience controls, including a secure authentication wrapper over a third-party API and a fallback image-generation strategy using AWS Bedrock to ensure reliability under peak load, while cutting perceived latency through video best-frame optimization.
Full Stack LLM Development Analyst — CV/ML Architecture & Pipeline Execution (POC) — Accenture
Bangalore, India | Oct 2025 – Dec 2025
- Built and validated end-to-end ML/CV training and inference pipelines — data ingestion, feature engineering, and spatio-temporal processing — with model debugging and accuracy checks to stabilize outputs across seasons.
Financial Operations Analyst Intern — HSBC
Gurugram, India | Jan 2025 – Jul 2025
- Led design and automation of regulatory reporting workflows, integrating data from multiple internal systems, and partnered with compliance and risk managers to translate regulatory requirements into scalable process solutions.
Risk Analyst Intern — Indialends
Gurugram, India | May 2024 – Jul 2024
- Automated risk and performance reporting pipelines for 1,300+ loans (saving 48+ hours annually) and built SQL-based validation checks to enhance loan-data integrity for reporting and compliance.
Projects
lakehouse-rag — Enterprise RAG with Access Control
Personal Project
- Built an enterprise RAG system over PDFs, a wiki, tickets, Slack, email and a live Postgres database, routing each question to hybrid search, guarded text-to-SQL or both, with per-caller access control enforced by row-level security and a CI gate of zero ACL leaks.
- Ingested through a Spark/Delta Lake medallion lakehouse on Databricks and ran it end to end on AWS (S3, SQS, Lambda, RDS, Bedrock) via Terraform: 0.917 answer correctness, 0.986 recall@k and $0.014 per 121-question evaluation run.
Research Summarizer using RAG + Multi-Agent LLMs
AI & Knowledge Management
- Built a RAG-based summarization pipeline for 1,000+ research papers, combining retrieval with sequential agents (CoT planner -> associative reasoning -> composer) and a Streamlit chatbot UI for grounded, interactive synthesis.
WhisperForge — GPU-Accelerated Speech-to-Text in Rust
Personal Project
- Built a high-performance Rust implementation of OpenAI's Whisper speech-to-text with GPU acceleration (Burn + WGPU; Vulkan/DX12/Metal), delivering SOTA decoding (beam search, temperature fallback, quality gates) on par with faster-whisper.
- Engineered per-token timestamps via cross-attention peaks (~100 ms precision) and speaker diarization, with INT8 quantization for ~4x model compression and real-time streaming transcription; shipped as a `wforge` CLI on crates.io.
Edge Embeddings + Rust/WASM Recommender
Personal Project
- Quantized and compressed embeddings (~4x smaller) into a Rust/WASM pipeline with MessagePack data bundles, delivering low-latency, in-browser song recommendations running entirely on the edge with no server round-trips.
qrdrop — Encrypted Peer-to-Peer File Transfer
Personal Project
- Built an end-to-end encrypted, serverless file-transfer tool: a QR code carries the key, ephemeral ECDH with per-chunk AES-GCM secures the WebRTC channel, and a four-emoji SAS check confirms the pairing; an air-gapped mode streams files as fountain-coded QR frames.
- Ran one protocol core across browser, Node CLI and a Tauri desktop/mobile app, shipped via npm, Homebrew, Scoop, the Microsoft Store and a self-hosted F-Droid repository with reproducible builds.
Skills & Tools
Generative AI & LLMs: RAG, Text-to-SQL, Multi-Agent Orchestration, Chain-of-Thought Reasoning, LangChain
Cloud & Platforms: AWS (Lambda, API Gateway, ECR, IAM, Bedrock, S3, SQS, RDS), Databricks, GCP (Vertex AI, Gemini)
Programming: Python, Rust, JavaScript (Node.js), CUDA C++, SQL (T-SQL, Spark SQL)
Data Engineering & Automation: Scalable pipelines, SQL-based data validation, Workflow automation, Robotic Process Automation
Business & Analytics: Requirements & PRDs, Source-to-Target Mapping, Corporate Strategy, Tableau
Tools: Terraform, Docker, Git, Jupyter
Education
MBA — Business Analytics — Birla Institute of Technology & Science (BITS) Pilani
2023–2025 | 87.3%
M.Sc. — Computational Biology — Indian Institute of Science (IISc) Bengaluru
2019–2023 | 74%
B.Sc. — Biological Sciences — Bangalore University
2016–2019 | 71.4%
Achievements
- AIR 24 in Joint CSIR-UGC NET (2019)
- AIR 64 in IIT JAM (2019)