Soham Dalal profile photo 1 on Tal

Soham Dalal

AI Native Engineer | GenAI & Agentic AI | Python, LLMs, RAG, MCP, LangChain, FastAPI, SQL, Spring Boot, Docker, AWS | Ex - AI developer @MavenArk

Mumbai

< 1 year of work experience

MavenArk
Jun 20265 mosPresent

Mumbai

Engineered a multi-threaded web crawler and document ingestion pipeline in Python (using Playwright and BeautifulSoup) that automated the retrieval of public financial disclosures and PDFs from company websites, the RBI, and the MCA registry. Developed an OCR-based table parser and glyph-repair engine to extract unstructured financial statements from scanned PDFs, mapping heterogeneous reporting standards (Indian GAAP and Ind-AS) into a unified MongoDB document store. Built automated cross-reconciliation and anomaly-detection modules that validated extracted financial tables against double-entry accounting rules (e.g., verifying Balance Sheet balancing equations) to ensure 100% mathematical integrity. Created and deployed a Model Context Protocol (MCP) server as an external gateway, enabling LLMs (ChatGPT, Claude, Gemini) to natively query the clean financial database with page-and-coordinate-level source provenance to prevent hallucinations. Designed and implemented full-stack interactive review screens in React (Vite) and FastAPI, displaying extracted data tables side-by-side with dynamic, pixel-cropped PDF source snippets to facilitate rapid manual corrections and auditability.
MavenArk
Jun 20263 mosAug 2026

Built an NBFC financial intelligence platform component using LLMs to extract balance sheet and P&L parameters from company filings. Validated extracted financial parameters against source PDF snippets to ensure data accuracy and reliability.
Indian Institute of Technology, Patna
May 20253 mosJul 2025

Patna

Dept. of CSE Under Dr. Joydeep Chandra (Associate Professor) Worked on Explainable Spatio-Temporal Graph Neural Networks (STGNNs) for detecting phishing scams on the Ethereum blockchain, combining GNNs with Explainable AI for interpretable cybersecurity solutions.
BSA Consultants
May 20253 mosJul 2025

Navi Mumbai

Developed a responsive corporate website using React, improving user engagement and navigation experience. Built reusable components and implemented REST API integration, enhancing scalability and functionality.
IISER Bhopal
May 20233 mosJul 2023

Bhopal

Soham’s AI Stack

langchain
langchain
langgraph
langgraph
AutoGen
AutoGen
Pinecone
Pinecone

Education

SVKM's NMIMS Mukesh Patel School of Technology Management & Engineering

Bachelor of Technology - BTech, Information Technology

SVKM's NMIMS Mukesh Patel School of Technology Management & Engineering

20234 yrs2027

High School Diploma, PCM

Moolji jaitha college

20212 yrs2023

Join Soham and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Soham Dalal on Tal

Headline

AI Native Engineer | GenAI & Agentic AI | Python, LLMs, RAG, MCP, LangChain, FastAPI, SQL, Spring Boot, Docker, AWS | Ex - AI developer @MavenArk

About

Recently completed a role as Python Developer at MavenArk, where responsibilities included building a multi-threaded web crawler and document ingestion pipeline utilizing Playwright and BeautifulSoup. Developed an OCR-based table parser to extract unstructured financial data and automated cross-reconciliation modules ensuring 99.99% mathematical accuracy. Currently pursuing a Bachelor of Technology in Information Technology from SVKM's NMIMS Mukesh Patel School of Technology Management &amp; Engineering. Core competencies include public speaking, Java programming, and Spring Boot, with a focus on applying technical skills to practical and innovative projects in backend web development and automation.

Location

Mumbai

AI stack

langchain

langgraph

AutoGen

Pinecone

Work experience

Python Developer at MavenArk, Jun 2026 - Present · 5 mos, Mumbai, Engineered a multi-threaded web crawler and document ingestion pipeline in Python (using Playwright and BeautifulSoup) that automated the retrieval of public financial disclosures and PDFs from company websites, the RBI, and the MCA registry. Developed an OCR-based table parser and glyph-repair engine to extract unstructured financial statements from scanned PDFs, mapping heterogeneous reporting standards (Indian GAAP and Ind-AS) into a unified MongoDB document store. Built automated cross-reconciliation and anomaly-detection modules that validated extracted financial tables against double-entry accounting rules (e.g., verifying Balance Sheet balancing equations) to ensure 100% mathematical integrity. Created and deployed a Model Context Protocol (MCP) server as an external gateway, enabling LLMs (ChatGPT, Claude, Gemini) to natively query the clean financial database with page-and-coordinate-level source provenance to prevent hallucinations. Designed and implemented full-stack interactive review screens in React (Vite) and FastAPI, displaying extracted data tables side-by-side with dynamic, pixel-cropped PDF source snippets to facilitate rapid manual corrections and auditability.

Software Engineer Intern at MavenArk, Jun 2026 - Aug 2026 · 3 mos, Built an NBFC financial intelligence platform component using LLMs to extract balance sheet and P&L parameters from company filings. Validated extracted financial parameters against source PDF snippets to ensure data accuracy and reliability.

Summer Research Intern at Indian Institute of Technology, Patna, May 2025 - Jul 2025 · 3 mos, Patna, Dept. of CSE Under Dr. Joydeep Chandra (Associate Professor) Worked on Explainable Spatio-Temporal Graph Neural Networks (STGNNs) for detecting phishing scams on the Ethereum blockchain, combining GNNs with Explainable AI for interpretable cybersecurity solutions.

Full Stack Developer at BSA Consultants, May 2025 - Jul 2025 · 3 mos, Navi Mumbai, Developed a responsive corporate website using React, improving user engagement and navigation experience. Built reusable components and implemented REST API integration, enhancing scalability and functionality.

Summer School 2023 at IISER Bhopal, May 2023 - Jul 2023 · 3 mos, Bhopal

Education

Bachelor of Technology - BTech, Information Technology, SVKM's NMIMS Mukesh Patel School of Technology Management & Engineering

High School Diploma, PCM, Moolji jaitha college