AI Engineer with 1+ years in Machine Learning & Fullstack Development
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Highly motivated AI Engineer with 1.1 years of experience in developing and deploying scalable AI solutions. Proficient in Python, PyTorch, and various machine learning frameworks, with a strong background in full-stack development. Demonstrated ability to build complex systems, from search engines to agentic lead-curation platforms, leveraging technologies like Elasticsearch, AWS, and Next.js, and contributing to open-source and technical writing.
Indian Institute of Technology Roorkee
B.Arch
N/A – June 30, 2027
Recepto.ai
ML Engineer Intern
October 1, 2025 – April 1, 2026
India
Orange Search
Co-Founder
June 1, 2025 – January 1, 2026
India
Sustainico
Full Stack Intern
May 1, 2025 – June 1, 2025
India
Tenyson: Multi-Step Unsloth Research Orchestration
June 1, 2025 – June 1, 2026
Runs multi-step Unsloth workflows (SFT, RL, eval) from two Python files: an experiment file that declares stages and imports dataset/rubric hooks, and a task file that defines dataset-generation functions and rubrics. Supports linear and parallel experiment chains while hiding training setup, cloud runners, HF adapters, W&B telemetry, and the usual Unsloth boilerplate. Updates the run report with the W&B link when a stage starts and stage metrics when it finishes; launch post gained traction on Twitter.
View ProjectGPT-2 From Scratch
June 1, 2025 – June 1, 2026
Built custom tokenizer, positional embeddings, multi-head self-attention, FFN, LayerNorm, and the training loop from scratch. Integrated inference and attention optimizations including KV caching, FlashAttention, and grouped-query attention.
View ProjectSynthetic RL Data at Scale for Small Models on Tau-Bench-3
June 1, 2025 – June 1, 2026
Benchmarking Qwen 4B on Tau3 Banking to beat the GPT-5.2 baseline (25%), first by building a custom harness and then generating synthetic RL data at scale. The tricky part is Tau3's 700+ knowledge docs, database state, and tools hidden inside the docs; next steps are distillation and RL on Qwen 4B.
View ProjectCultural Fit Analysis
The candidate's project diversity, ranging from deep learning model implementation (GPT-2 From Scratch) to large-scale data systems (Orange Search) and agentic AI (Recepto.ai), indicates a broad interest and adaptability. Their involvement in a student developer club suggests a collaborative and community-oriented mindset. The target role of 'AI Engineer' aligns well with their demonstrated skills in RL, Transformers, and AI Agents. The breadth of skills across ML, MLOps, and full-stack development suggests a versatile individual who can contribute to various aspects of an AI product lifecycle.
Soft Skills & Operational Fit
The candidate demonstrates strong initiative and problem-solving skills through their co-founder role and personal projects. Their experience in rebuilding an agentic system from first principles suggests a critical thinking approach. The Google Developer Student Club involvement indicates teamwork and community engagement. However, the resume does not provide explicit details on stress handling or direct team collaboration within a professional ML engineering context, which would require further validation.