PeopleofSuperintelligence

Builder

Saiful Haq

Saiful Haq is an AI researcher and engineer whose work spans enterprise search, language models, multilingual retrieval and document AI. He pursued PhD studies at IIT Bombay from 2022 to 2026 and joined Hyperbots’ founding team in October 2023. He coauthored Angel and DSPy, was first author of IndicIRSuite for 11 Indian languages, and studied mental-set effects on LLM reasoning. In 2026, he coauthored DocuLite and SAVIOR for invoice extraction and financial OCR. As of October 2026, he is CTO at Asicbot AI.

Now at
Hyperbots Inc. and Asicbot AI
Works
12
Badges
Links
Google Scholar and LinkedIn

Publication numbers from OpenAlex, as of 22 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.

Timeline

  1. 2022–2026

    PhD studies at IIT Bombay

    He pursues PhD studies at IIT Bombay from 2022 to 2026.

  2. October 2023–March 2026

    Hyperbots founding team

    He joins Hyperbots as an AI/ML Research Engineer on its founding team in October 2023. From October 2024 to August 2025, he is Lead AI/ML Research Engineer. From August 2025 to March 2026, he is Director of Artificial Intelligence and Staff Research Engineer, and he leads AI, Product & Engineering for HyperAPI from January 2026 to March 2026.

  3. 2023

    Angel for nonprofit search

    In 2023, he coauthors Angel with Ashutosh Sharma and Pushpak Bhattacharyya, a paper in the EMNLP Industry Track proceedings that builds nonprofit enterprise search. It adapts ColBERT with syntax-aware local attention and dense pseudo-relevance feedback, improving MAP@10 by 0.14 and MRR@10 by 0.16 over the state-of-the-art baseline.

  4. 2023–2024

    The DSPy programming model

    He is one of 13 authors of DSPy, submitted to arXiv in October 2023 with Omar Khattab as first author and Matei Zaharia and Christopher Potts among the coauthors; ICLR lists the paper in 2024. Its programming model abstracts language-model pipelines as text transformation graphs, and its compiler optimizes a pipeline to maximize a given metric.

  5. 2024

    Retrieval models for Indian languages

    He is first author of IndicIRSuite with Ashutosh Sharma, Omar Khattab, Niyati Chhaya and Pushpak Bhattacharyya, published in the ACL short-paper proceedings in August 2024. It provides INDIC-MARCO, a machine-translated MS MARCO dataset in 11 Indian languages, and 11 monolingual Indic-ColBERT retrieval models.

  6. January 2025

    Mental sets in LLM reasoning

    In January 2025, he is first author of a study with Niyati Chhaya, Piyush Pandey and Pushpak Bhattacharya on mental sets in LLM reasoning. It compares Llama-3.1-8B-Instruct, Llama-3.1-70B-Instruct and GPT-4o under mental-set conditions. The paper defines a mental set as persistence with previously successful strategies.

  7. 2026

    Invoice extraction and OCR papers

    In 2026, he is first-listed author of DocuLite, a full paper accepted at the 4th Deployable AI Workshop at AAAI, with coauthors from Hyperbots Inc. and IIT Bombay. It adapts large language and vision-language models for invoice extraction using InvoicePy for synthetic text and TemplatePy for synthetic invoice templates, without human-annotated data. He also coauthors SAVIOR, a 2026 WACV workshop paper on sample-efficient adaptation of vision-language models for OCR representation.

  8. As of October 2026

    CTO at Asicbot AI

    As of October 2026, he is CTO at Asicbot AI, which builds AI agents across the chip development lifecycle.

Career

  1. NowHyperbots Inc.Director of Artificial Intelligence and Staff Research Engineer
  2. NowAsicbot AICTO

Education

  1. 2026IIT BombayPhD

From the biography.

  1. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes IIT Bombay. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  2. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes IIT Bombay. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  3. The median h-index is 48 at universities and 14 at frontier AI labsIncludes IIT Bombay. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.
  4. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes IIT Bombay. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Saiful Haq yet. Send us a free-licensed photo.

  1. Saiful Haq — Lacuna lacuna.tiptreesystems.com. Checked 21 October 2026.
  2. Saiful Haq - Ai/ml Research Engineer (founding Member) at Hyperbots Inc. | The Org theorg.com. Checked 21 October 2026.
  3. Saiful Haq - Co-Founder | Stackforce stackforce.co. Checked 21 October 2026.
  4. Angel: Enterprise Search System for the Non-Profit Industry - ACL Anthology aclanthology.org. Checked 21 October 2026.
  5. DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines (arXiv:2310.03714) arxiv.org. Checked 21 October 2026.
  6. IndicIRSuite: Multilingual Dataset and Neural Information Models for Indian Languages - ACL Anthology aclanthology.org. Checked 21 October 2026.
  7. Is your LLM trapped in a Mental Set? Investigative study on how mental sets affect the reasoning capabilities of LLMs (arXiv:2501.11833) arxiv.org. Checked 21 October 2026.
  8. Doculite Research | Hyperbots hyperbots.com. Checked 21 October 2026.
  9. SAVIOR: Sample-efficient Adaptation of Vision-Language Models for OCR Representation - WACV 2026 Open Access Repository openaccess.thecvf.com. Checked 21 October 2026.
  10. Asicbot AI - Chip Development, Reimagined asicbot.ai. Checked 21 October 2026.