Software Engineer Backend

Mridul Jain

Software Engineer

+91-788******* **************@*****.*** Linkedin

Profile

Backend Engineer with 3 years of experience building scalable, backend systems, GenAI-powered extraction workflows, and RESTful APIs in Python. Skilled in developing low-latency, multi-threaded and multi-processing systems with FastAPI, Tornado, and PostgreSQL, with a strong emphasis on microservice architecture, data quality, observability, and performance benchmarking. Adept at solving complex backend, GenAI, and data workflow challenges through root cause analysis and system profiling in fast-paced, product-driven environments. Experience

Alphastream.ai August 2023 - Present

Software Engineer Bengaluru, India

• Pipeline Reliability Debugging: Built and maintained a fault detection framework to pinpoint issues across ingestion, transformation, and validation stages. Implemented advanced text matching and paragraph alignment techniques to compare pipeline outputs, storing intermediate results in AWS S3 for scalable access and auditability. This enabled faster fault isolation, reduced debugging time by 40%, and improved data quality for downstream systems.

• GenAI-Powered Extraction: Contributed to a large-scale pipeline extracting 1000+ datapoints from 500-page legal PDFs. Built the Investment Covenants DP model using multi-label classification to identify 28 datapoints from classified sections. Integrated a hybrid workflow where extracted paragraphs were passed to paid LLM APIs and fine-tuned open-source LLMs for value extraction. For 6 datapoints with limited training data, designed a bulk prompting strategy to ensure accurate retrieval under low-data constraints.

• Data Access Integration APIs: Designed and implemented RESTful APIs to enable efficient, on-demand retrieval, transformation, and validation of data from relational databases. Supported parameterized queries for column-level selection and dynamic filters, streamlining data extraction workflows for analysts and data scientists and reducing manual data preparation time.

• Cross-Product Benchmarking System: Developed a universal benchmarking framework to evaluate pipeline updates across products and releases. Captured module-level processing times and compared value and coordinate extraction accuracy with curated baseline data. Enabled fine-grained performance tracking and achieved a 3.5 speedup in statistical reporting through multi-processing.

• Production-Grade Data Validation Framework: Engineered a low-latency validation layer that integrates rule-based logic with historical data profiling to ensure clean, accurate ingestion. Achieved 8–9% increase in F1 score, reduced false positives by 90%, and accelerated validation by 40% through multi-threaded. Alphastream.ai August 2022 - August 2023

Software Engineering Intern Bengaluru, India

• Form Data Extraction Tool: Built a robust pipeline using PyMuPDF and Textract to extract key-value pairs from 1–2 page PDFs (both scanned and digital Acrobat formats), supporting 4 distinct form templates. Engineered modular parsing logic and integrated OCR fallback, achieving 90%+ field-level accuracy across documents.

Technical Skills

Programming Languages: Python, SQL

Frameworks & Libraries : FastAPI, Tornado, Pandas, NumPy Technologies & Tools : Docker, PostgreSQL, Redis, Jenkins, Git, Linux Cloud Platforms : AWS (S3, CloudWatch, Lambda, Textract) Leadership : Technical Ownership, Cross-functional Collaboration Education

Lovely Professional University Phagwara, Punjab

B.Tech, Computer Science and Engineering — CGPA: 8.09/10.0 July 2019 - July 2023